Você está na página 1de 102

Proceedings

of the
Second Australian
Acid Mine Drainage
Workshop
Editors: N.J. Grunden and L.C. Bell
Charters Towers, Queensland
28- 31 March 1995
Jointly organised by:
AUSTRALIAN CENTRE FOR
MINESITE REHABILITATION RESEARCH
P.O. Box 883
KENMORE, Old 4069
Published by:
MINERALS COUNCIL
OF AUSTRALIA
P.O. Box 363
DICKSON, ACT 2610
Australian Centre tor
Minesite Rehabilitation Research
( i)
Published by the Australian Centre for Minesite Rehabilitation Research.
ISBN: 0 646 24771 9
July, 1995
Copyright 1995, Australian Centre for Minesite Rehabilitation Research.
Articles in this publication are protected by copyright. Material from the publication may be
used providing both the author and publisher are acknowledged.
Individual papers are the responsibility of the author(s). The opinions expressed do not
necessarily represent the views of the publishers.
Enquires and requests for copies should be directed to:
The Director,
Australian Centre for Minesite Rehabilitation Research
P.O. Box 883
KENMORE, Old 4069
Australia
Telephone:
Facsimile:
(07) 3212 4555
(07) 3212 4574
Citation of this publication should take to form:
Grundon N.J. and Bell, L.C. (Eds.) (1995). Proceedings of the Second Australian Acid Mine
Drainage Workshop, Charters Towers, Queensland. 28-31 March 1995. (Australian
Centre for Minesite Rehabilitation Research: Brisbane, Australia).
Individual chapters should be cited:
Author(s) Name(s). (1995). Title of paper. In 'Proceedings of the Second Australian Acid
Mine Drainage Workshop', Charters Towers, Queensland. 28-31 March 1995. (Eds.
N.J. Grundon and L.C. Bell.) pp. 00-000. (Australian Centre for Minesite Rehabilitation
Research: Brisbane, Australia.)
Acknowledgments
The organising committee of Gavin Murray (Chair), Matthew Orr, Peter Ryan and Clive Bell
wish to thank Chris Palmer for his assistance with local arrangements, and Fay Gallie for her
administrative support.
Appreciation is also expressed to Mt. Leyshon Gold Mines Limited, Battle Mountain (Aust.)
Inc., Pancontinental Resources, Placer Pacific Limited, and Woodward Clyde for their
sponsorship.
Printed by The University of Queensland, Brisbane.
(ii)
TABLE OF CONTENTS
Welcome and Opening Address
Introduction to the Acid Mine Drainage
Workshop L.C. Bell
Session 1: The AMD Problem
Defining the AMD problem:
I. A corporate perspective
Defining the AMD problem:
II. An operator's perspective
Session 2: Planning for AMD
G.S.C. Murray, J.D. Robertson and
K.O. Ferguson
B. Price
Sampling and waste block modelling G.M. Johnston, G. Eastwood,
for acid producing waste rock at Cadia P. Scott, C. Joyce and P. Ryan
Advances in the prediction of acid
generating mine waste materials S. Miller and J. Jeffery
A manager's operational perspective
of AMD management at the Pajingo
Gold Mine /.A. Tredinnick and P.J. Cornwell ..... .
Session 3: Management Strategies
Practical engineering options to
minimise AMD potential A. Watson
Control of acid mine drainage through
water management at Mt. Leyshon
Gold Mine M.S. Orr
Session 4: Management Strategies (cont.)
Management of AMD at Woodlawn
Mines S.E. Tarlinton
Management of AMD at the Pine
Creek Mine M.N.R Fawcett
Development of a waste dump
management strategy at Mt. Leyshon
Gold Mine M.S. Orr
Session 5: Monitoring and Modelling Strategies
Application of oxidation rates in
rehabilitation design
Geochemical indicators of sulfide
oxidation and acid generation in the
field
A/. M. Ritchie
S. Miller
(iii)
Page
3
17
21
33
43
53
67
75
83
91
101
117
Session 6: Treatment and Control Strategies
Water flow in mine waste dumps and
their implications for AMD control L.M. Gilbert
Wetlands to treat AMD - facts and
fallacies D.R. Jones and B.M. Chapman ......... .
Assessment of protective covers and
dump behaviour G. W. Wilson
Session 7: Performance Evaluation and Risk Assessment
Financial and long term liability
associated with AMD
G. S.C. Murray, K.D. Ferguson and
C.H. Brehaut
Session 8: Future Technology Transfer and Research Needs
Assessment of the Rum Jungle
Strategy for acid mine drainage
control
Group Discussion Session
List of attendees
J. W. Bennett and M.D. Lawton ......... .
(iv)
Page
121
127
147
165
179
191
193
I
i
I
I
f
l
I
!
I
I
I
Second Australian Acid Mme Oramage Workshop {Eds N.J Grundon and L.C. Bell), pp 1 ~ 2
INTRODUCTION TO THE ACID MINE DRAINAGE WORKSHOP
L.C. Bell
Australian Centre for Minesite Rehabilitation Research, P.O. Box 883, Kenmore, Old 4069.
On behalf of the organising Committee, I would like to extend to you a warm welcome to this
Acid Mine Drainage (AMD) Workshop which has been organised jointly by the Australian Mining
Industry Council (AMIC)' and the Australian Centre for Minesite Rehabilitation Research (ACMRR).
I beliwe the response that we have had to this workshop is indicative o.f the increasing interest
in the topic of AMD in Australia, and the fact that we have representatives here from Papua
New Guinea, Indonesia and New Zealand indicates that AMD is also an issue in our near
neighbours. Indeed, it is fair to say that the prevention and remediation of AMD is of a major
concern for metalliferous and coal mining operations world-wide, and we are also pleased to
have Professor Ward Wilson from Canada with us to share his experiences on the topic.
This workshop is the second of what might reasonably be called a national workshop on AMD.
The first was held at Strahan in Tasmania in 1992. That workshop, which was attended by
about 30 persons, was organised by the Tasmanian Chamber of Mines. Subsequently, the
Environment Committee of AMIC established a sub-committee to identify priorities for environmental
research for the mining industry. One of the priority areas identified was AMD. An Acid Drainage
Working Group was then established to identify specific research and technology transfer issues
that needed to be addressed in the future. This group consists of Gavin Murray (Chairperson)
{Placer Pacific Limited), Alex Armstrong (Area Coal Australia Inc.), John Bywater (Old. Dept.
of Minerals & Energy), Peter Chandler (AMIRA), Geoff Farnell (EPA- formerly with AMI C), Alan
Hayter (Renison Goldfields Consolidated Limited), David Jones, (CSIRO), Tony McGill (N.T.
Dept. of Mines & Energy), Graeme Mcilveen (Normandy Poseidon Limited), Stuart Miller
(Environmental Geochemistry International), Matthew Orr (Normandy Poseidon Limited), ian
Ritchie (ANSTO), and Peter Ryan (Woodward Clyde).
At a meeting of the group on 22 March 1994, it was recommended that there was a need for
a second workshop to follow on from the Strahan workshop. An organising committee consisting
of Gavin Murray (Chairperson), Matthew Orr, Peter Ryan, and myself was established to organise
the Charters Towers workshop.
The objectives of the workshop are:
(1) Review the current state of knowledge on (i) AMD development and associated
liability, and (ii) strategies for its prevention and remediation;
(2) Assess practical management strategies for AMD through sites visits; and
(3) Determine future technology transfer and research needs.
Now called the Minerals Council of Australia
BELL
Welcoming Address.
In determining the program for the workshop, the organising committee sought the advice of
all members of the AMIC AMD Working Group who provided constructive comment concerning
both content and speakers.
The Workshop has sessions covering (1) the AMD Problem, (2) Planning for AMD, (3) Management
Strategies, (4) Monitoring and Modelling Strategies, {5) Treatment and Control Strategies, (6)
Perfonmance Evaluation and Risk Assessment. and (7) Future Technology Transfer and Research
Needs with the latter session dominated by general discussion. Complementing the fonmal sessions
are visits to the Pajingo Mine (Battle Mountain (Aust.)), MI. Leyshon Gold Mine (MI. Leyshon
Gold Mines Limited) and Thalanga Mine (Pancontinental Resources).
There are many places where a workshop of this nature could have been held. While larger
towns would have had venues to cater for larger numbers, it was felt that Charters Towers had
the advantages of (1) being close to a number of mining operations enabling site visits each
day, and (2) the mines in the area are using a number of innovative practices which would be
of interest to a wider audience.
The organising committee believed that it was important that this workshop be a true workshop
with the maximum opportunity for discussion of the various issues. Thus the speakers will be
restricted to 20 to 25 minutes allowing 15 to 20 minutes for discussion after each paper. On
the last day, we have set aside an additional 2 hours for general discussion.
Because of the overwhelming response to the workshop, we have had to increase the number
of registrants over the 55 that we had originally planned for. I am delighted to be able to indicate
that we have 78 registrants at the workshop. Unfortunately many more wanted to attend, but
the constraints of the venue and transport at the mine-sites meant that the organising committee
had to set this as the upper limit.
At this workshop, we have a good representation from each of the industry, government, research
and consulting sectors. Additionally we have a balanced mix of formal presentations and mine-site
visits and a considerable period set aside for discussion. I am confident the workshop will be
a successful one!
BELL
2
S;:,cond Austral1an Ac1d M1ne Dra1nage Workshop (Eds N J Grundon and L C Bell), pp 3-15
DEFINING THE AMD PROBLEM I.
A CORPORATE PERSPECTIVE
G.S.C. MurrayA, J.D. Robertson
8
, and K.D. Fergusonc
APiacer Pacific Limited, G.P.O. Box 4315, Sydney, NSW, 2001.
8
Piacer Dome Inc, P.O. Box 49330 Bentall Postal Stn, Vancouver, Canada V7X 1 P1.
0
Piacer Dome Canada Ltd, P .0. Box 49305 Ben tall Postal Sin, Vancouver, Canada V7X 1 P1.
ABSTRACT
Placer Dome Incorporated (POl), a precious and base metal mining company, recognises acid
drainage (AMD or ARD) as one of the most significant environmental issues facing its activities
around the world. While the major variables involved in acid generation are known, they still
cannot be fully quantified and many of these variables remain the focus of various international
research. In response to the significance of this issue to the corporation, POl has initiated an
aggressive program to assess the ARD potential of all of its operations and activities worldwide.
To date, studies to rank the potential significance of the problem have been undertaken at more
than 20 of the company's operations and projects. The rationale and mechanics of the PO/
assessment program are outlined with the objective being the application of predictive information
to prevention and/or control strategies on a site-specific basis. Research programs are conducted
and supported on an ongoing basis as an integral component of the program.
1.0 INTRODUCTION
One of the most significant environmental issues facing the mining industry is acid drainage
(AMD or ARD, as used in this paper) which affects all sectors of the industry including coal,
precious metals, base metals, and uranium. Although this issue is not new and has an extensive
history spanning decades and even centuries in Europe, it is still not fully understood. While
the major contribution variables are known, they cannot be fully quantified. Many of these variables
are still the subject of research by various task forces and international research groups.
Empirical knowledge is being acquired to predict and prevent ARD, but the lack of a full
understanding of ARD dynamics has caused concerns among the regulatory agencies. This
is particularly reflected in the permitting and approval procedures associated with new mines.
The result has been the demand by agencies for overly-conservative control strategies and
significant performance bonds in an attempt to provide assurances that any ARD problems can
be managed, regardless of the risk.
Placer Dome Incorporated (POl) is a precious and base metal mining company based in Canada
with major operating subsidiaries in Australia, Canada, South America, and the USA. In Australia,
the operating .subsidiary Placer Pacific Limited, manages six operating mines and exploration
within Australia and the Asia-Pacific region. Placer Pacific Limited is 75% owned by PDI and
adopts a similar operating philosophy and management strategy to PD I. This paper provides
an overview of the rationale and the mechanics of the POl ARD assessment program which
was developed to establish a sound and standard corporate approach.
3
MURRAY, ROBERTSON & FERGUSON
Defining the AMD Problem I.
2.0 ASSESSMENT AND MANAGEMENT STRATEGY
Acid rock drainage (ARD) is the term used in North America to refer to problems associated
with acid drainage, while acid mine drainage (AMD) remains in common usage in Australia.
PDI has adopted ARD as the standard reference to refer to any issues related to the oxidation
of sulfide minerals exposed by mining or exploration which may include issues associated directly
with rehabilitation or water quality. Experience to date has shown that ARD issues may involve
acid drainage from waste rock dumps and tailing residue emplacements (impacting on downstream
water quality), development of acid conditions in exposed surface materials (potentially affecting
rehabilitation), increased solubility and/or release of metals (irrespective of actual pH) and increased
salinity or solute loads (oxidation and neutralisation products).
The POl Board of Directors have approved an environmental policy and associated strategic
plans which require a thorough knowledge of the ARD risks for all of the company's activities.
Also required is the implementation of appropriate controls to minimise potential environmental
impacts during operations and after closure. In order to address these concerns in the best
practical manner, POl has initiated and conducted over the past five years comprehensive ARD
reviews of each minesite and advanced exploration projects. The acquired information forms
the basis for internal environmental operating plans and is also a key part of the rehabilitation
plans for all minesites. Some sites have required detailed investigations for more complete
assessment, and research projects have been established at a number of sites.
2.1 Corporate Philosophy
POl's mines have been classified into four general categories: future mines (advanced exploration
projects/acquisitions), newly designed/constructed mines, existing/operating mines, and
decommissioned/rehabilitated mines, for the assessment of ARD potential. While there are
different philosophies and strategies developed for each group, there is a set of underlying principles
which apply generically to ARD prediction as outlined below:
(1) Each minesite is unique with respect to geology, climate, and site configuration
and, hence, the application of rigorous routine prescriptive methods is not appropriate.
(2) Sulfide mineral reactivity is highly variable and ARO may not occur at all, or be
delayed for several years after exposure from mining. Consequently, it is never
too late to assess the ARO potential at existing mines using predictive methods
with additional integration of empirical site information.
(3) Prediction and prevention of ARO is the most desired and appropriate control plan
for all sites.
( 4) The results of the predictive assessments may justify changes in mining operations
and waste rock and tailing disposal practices to reduce the potential for ARO.
(5) The application of predictive methods at an early stage of an advanced exploration
project will allow a phased approach for further studies and, if necessary, extensive
special planning of waste dumps and tailing dams.
(6) Internal training of PDI technical staff in the prediction, assessment, and control
methods for ARO is an integral component of proper ARD management.
MURRAY, ROBERTSON & FERGUSON
4
Defining the AMD Problem I.
(7) Ongoing applied research is needed at sites with ARO potential to track the success
of the implemented prevention strategies and to augment the knowledge base
for the development of improved techniques within the minesite and POL
(8) Continued and significant support is required for generic research through industry
initiatives and at individual POl sites to allow a better understanding of the processes
associated with ARD.
2.2 Program Approach
The POl approach to ARD assessments for its new and existing mines is illustrated as a flow
sheet in Fig. 1. This approach has evolved during the assessment and management of ARD
potential at more than 20 of POl's operations and projects worldwide. A number of key decision
points are included in the process which dictate the direction and nature of each phase of the
assessment program.
Each site-specific program consists of a team approach composed of several key players and
facilities. These personnel vary with the classification of the project, but the common denominator
is a joint effort between corporate personnel (environmental and engineering), specialised experts,
and the exploration, design and operating personnel who will eventually apply the end product
to the design or operation. An experienced site geologist and mine design engineer usually
assist the environmental personnel in an initial overview assessment and provide continuity
for ongoing work.
Each program is phased over several stages as shown in Fig. 1. The initial overview is conducted
on-site with the study team, and a plan is developed for compiling the existing information and
for collecting additional data and samples for assessment, as required. Ongoing site meetings
are held as needed after the initial data has been collated and assessed. Interim reports are
prepared by experienced internal personnel or external specialists depending on the level of
expertise and/or detail required.
Ongoing technical support is provided for each project depending on the techn.ical strengths
and capability of the site-based technical personnel and the magnitude of the potential problems.
Through this process, a rnore complete awareness of ARO is developed throughout the operating
workforce, and technical skills are developed by a variety of internal POl personnel. This achieves
self-sufficiency for each independent operation and results in less dependence on the limited
resources of the corporate technical group. However, this approach still provides continuity
across the full spectrum of POl operations and allows an exchange of information between sites.
Furthermore, the broadened experience gained by personnel through job transfers results in
a better appreciation of subsequent ARO issues.
2.3 Testing Program
The POl standard approach is to use both static (acid base accounting) and kinetic testing for
predictive assessment. The acid base account (ABA) is the balance between the theoretical
acid generating potential from sulfide oxidation and the inherent acid consuming ability of the
material. POl has also adopted the North American convention in terminology and units for
expressing the results of testing, which are used in this paper. The North American and Australian
conventions are compared below for general reference.
5
MURRAY, ROBERTSON & FERGUSON
Defining the AMD Problem I.
Australia
North America
Acid Producing Potential (APP) Acid Potential (AP)
(both usually calculated from total sulfur or sulfide %)
Acid Neutralising Capacity (ANC) Neutralisation Potential (NP)
(both usually determined by acid addition and back titration)
Net Acid Producing Potential (NAPP)
(NAPP = APP-ANC)
Net Neutralisation Potential (NNP)
(NNP = NP-AP)
Classification of potentially acid generating materials
positive (+Ve} NAPP
negative (-ve) NNP
Standard units for expressing ARD potential
kg H2S04 t ' kg (t) CaC0
3
t-' (1 ,000 C')
(Note: 1 kg H
2
S0
4
t-' = 0.98 kg CaC0
3
t-')
While static and kinetic testing to assess the ARD potential is available from commercial
laboratories, POl has developed the capability to perform these tests on a routine basis in its
own research laboratory in Vancouver which allows continuity, better integration, rapid
implementation of additional assessment work where needed, and a standard application of
techniques. Quality assurance and quality control (QAJQC) programs are also run for each
project using external commercial laboratories for comparative analysis.
The number of samples subjected to static testing ranges from 50 to 2000, depending on the
objective and status of the program. If spatial modelling of an orebody with ABA data is warranted,
several thousand analyses may be performed in a phased approach to provide a good cross
section of the rock and mineralisation types within the deposit.
A number of kinetic testing methods are available with the standard 'humidity cell' procedure
requiring a minimum duration of 20 weeks, while most column tests are run for longer durations.
However, this time frame has been shown to be inadequate for most samples unless they are
extremely high in sulfur content, low in buffering capacity and/or highly reactive. POl programs
involving kinetic testing typically continue for a much longer time frame which may extend up
to two to three years depending on the progressive results. This strategy allows more flexibility
and a complete assessment of highly buffered and marginally reactive materials.
Consultants specialising in ARD assessment and management are utilised as required to assist
in collating and interpreting data or to provide an independent review of in-house assessments.
External specialists are also contracted for the evaluation of rock samples and thin sections
with respect to general mineralogy and the examination of oxidation of mineral surfaces. Important
correlations between mineralogy and static and kinetic test results have been discovered which
aid interpretations.
MURRAY, ROBERTSON & FERGUSON
6
"T1
tEl
"U
iii"
()

0
0
3
<D
:I
0
Q
"0
0
il1
ro
a.
Dl
"0
"0
0
!),)
0
:::T
0
Dl
0
a:
0
0
:><;"
a.

:I
Dl
(Q
CD
Dl
"'
"'
CD
g:
3
CD

2.
:I
CD

Dl
:I
a.
CD
X
ii
5"
(Q
3
:I
CD
!'>
Defrning the AMD Problem I.
..
.--- .--
r---.--_..., b> ;51
.-
a I I I {1)
- (f) w
(/) ..,_,Ol:; "0 (/)
5) r-
<D :>(f)'" -u OJPl = I
(Q ...- o.ooOi"..._ )>ii: o -
'<
0 e!.o.3 :; U> I
<D =cu<O m'" = ,-J
a5 c.6 rocn --=:
0 en cE -a
-- 0
0. (Q
..-' p; r---' .-- .--- .--
... 3 "U ::0
--' ::0 55 ;51 <D ,--
en (f) Ol :s.
.-A CD (f) CD m
"UP> 1\) <D ::;: (f)
.-- g.6. -z;,
)> g}of : )>
::0 =:::!. " ()" ()" 03 ::J"
0 rol rol ... iii' ._
,-1 g-
ifil ::05' (Q 2: -
ro (f) :::> Om "U -u o rol
3 +- 1 ... 3' .- ,g 0 3
gj () :::!. 0 u;
- ::J" :J (/)0 PJ PJ - - _j
(f) (Q "3. 3 3 g --
=::- .-J r m::; _ _j
<D3 - '--'
<D (Q PJ
(Q
'< og ...__. 5
_J
<DO
0:2. --'
0
a.
(Q

--+---------
.-- } r
"U )>
co /\ o.
1..-- 0 =l 0. <D ..____ " :; g
3 ffi.rol[SJ - g
:::3
._, ro ro
2 --'
.- g-
:J r-- -o co
-=?5" ,g 3
o W::l Oo -.. co
:::> '<<O I::;: ;4 PJ
3
'--' L-J
----'

7
MURRAY, ROBERTSON & FERGUSON
Defining the AMD Problem I.
Table 1.
Summary of acid drainage assessment variables and Information availability.
Source variable
Project classification
Explor. Design Operating Decomm.
Regional information:
Groundwater quality data

Surface water quality data

Exploration/mine water/pit wall seepage water quality

Exploration database:
Total sulfur analysis and/or ABA on all samples

Drill core visual estimates of carbonates, sulfides etc

Petrologyfmineralogy information available

Samples for predictive testing:
Weathered/oxidised surface rock samples

Pit wall rock/underground chip samples

Drill core samples for static testing (ABA)

Drill core samples for kinetic testing

Bulk samples for small scale field tests (barrels)

Bulk samples for large scale field tests (test dumps)

Metallurgical testing samples (tailing residues etc)

Extensive drill core data for spatial modelling

Site facilities:
Existing waste rock dumps

Existing tailings dams
Other local historical mine facilities for review

Iron staining: seepages; pit walls, tailing, etc

Site design information:
Climate data (precipitation/evaporation)

Site hydrology (water balances)


Special waste dump designs developed
+

Special tailing design developed

Seepage containment strategies exist

Site runoff and drainage strategies exist


Unavailable,
Possibly available, + Likely available, + + + available.
MURRAY, ROBERTSON & FERGUSON
8
Defining the AMD Problem I.
2.4 Assessment Program
Each of the four operational categories is assessed in a slightly different manner depending
on the available information. In general, the assessment consists of the compilation and integration
of a whole range of information to provide a general order of magnitude overview of the ARD
issue. Additional work is defined to clarify some areas and to refine others. In some cases,
specific projects are initiated to provide short and long term information on specific issues. Table
1 presents a matrix which summarises some of the assessment variables and indicates the
availability of information in each case.
The results from all of these studies are summarised in several formats for presentation to a
wide range of end users including the mine operators, consultants and PDI management. A
description of the form and objective of each of these formats follows.
2.4.1 Status summary table
Progress on the ARD assessment program is presented in a status summary table (Table 2)
which is designed to progressively illustrate both the information acquired and the relative risk
of ARD potential. This table also reflects the status of the implementation of prevention and
control strategies if ARDis predicted. As the control strategies become more complex through
design and implementation, the project maybe moved down a level in the general risk category.
However, some sites will always be a higher risk, and the main focus is to ensure preventive
action and systems are operational and effective .
Table 2. Example of ARD assessment status summary table.
Operational Issue Ranking Programs 1994 Status/Activities
Activity ;
2 3 4
Mine A
Mine B
Mine C
MineD
Mine E
Mine F
Mine G
Mine H
Mine I
Key:
Issue Ranking:



1 Low cost or concern
through to
4 High cost or concern"



Report Testing
1993 No Report completed
1994 Yes Continuing kinetic testing program
1993 Yes Report completed +management plan
1993194 No Report completed
1992 Yes Continuing testing program
1993/94 Research on covers continuing
1993 Yes Revised closure costs+ research plan
1994 Data reviewed
1994 Program initiated
Program Ranking:
Low knowledge or control
through to
+ + High knowledge or control.
Overview report done
Detailed report done or low risk
+ + Operating plan developed
+ + Plan verified by practice
9
MURRAY, ROBERTSON & FERGUSON
Defining the AMD Problem I.
2.4.2 ABA presentation and analysis
There is much debate over the interpretation of ABA results for single samples in the context
of mixed waste rock dumps composed partly of acid and non-acid producing materials. Graphical
techniques to visualise the data and to compare them to literature reported criteria are important
tools for assessing the probability of ARD. However, it is important to note the limitations of
ABA and apply it conservatively as a screening tool until correlated with known field behaviour.
The database compiled by POl for all of its projects, as well as other known ARD sites around
the world, provide relative data sets to assist in the interpretation of ABA data from each new
project. Within POl, the ABA result is expressed as the 'Net Neutralisation Potential' (NNP) with
negative values indicating the potential for a sample to generate acid. A statistical approach
to the assessment of ABA data is generally applied to quantify the potential risk of ARD. This
involves preparing a frequency distribution of the NNP data obtained which also enables an
evaluation of the adequacy of existing sample selection.
An example of an NNP frequency distribution for combined and separated flotation tailing streams
is presented in Fig. 2. This figure indicates that the 'rougher' tailing have a lower risk of ARD
than the 'cleaner' tailing and that the 'combined' stream produces a material which is likely to
be classified as potentially acid generating. Based on the statistics of the distribution, a probability
of exceedance plot can be generated which provides a more refined tool for assessing the
probability of ARD.
1::
.2
:;
.0
....
o;
c
>.
u
1::
Q)
:l
0'
Q)
...
u..
0.6
0.5
0.4
0.3
0.2
0.1
Legend
Cleaner Tailing
Rougher Tailing
Combined Tailing
\./"/.._
0.0 = 'I - '.1 I "-J
250 -150 -50 0 50 150 250
Net Neutralisation Potential (kg CaCOJ!t)
Fig. 2. Relative Net Neutralisation Potential for flotation tailing
streams.
2.4.3 Acid generating potential unity line
A method of graphically presenting ABA data and statistics, adapted from Miller et al. (1991)
where AP is plotted against NP, is also used to assess sample distribution and the relative risk
of ARD. An important component of these ABA plots, which are illustrated in Figs. 3 and 4,
is the unity line ( N N P ~ O ) or the line where Neutralisation Potential equals Acid Potential. This
line represents an NP/AP ratio of 1 with data points above the line having the potential to generate
MURRAY, ROBERTSON & FERGUSON 10
Defining the AMD Problem I.
acid, while samples below the line have adequate neutralising potential to counteract the potential
acid production. This assessment is performed on all materials examined, including tailing residues
and individual rock types.
~
250
I POTENTIALLY: ACID GENERATfG I
0 200 .......--1
()
m
()
Ol 150
6
~ 100
1::
Ql
0
c..
50
"0
'(j
<(
0
NON ACID GENERATING
0 50 100 150
Neutralisation Potential (kg Caco
3
tt)
Fig. 3. Acid-base account plot for tailing composites.
200
To facilitate an assessment of the ARD potential, each ABA data set is plotted with their statistical
limits. Comparison of existing data blocks with new data provides a quick guide to the general
portion of the samples with ARD potential and also the order of magnitude of potential problems
relative to existing sites. Figure 3 illustrates the application of this technique to tailing samples
for a proposed operation with 95% confidence interval ellipse indicating that the tailing will be
potentially acid generating .
Data for four known acid-generating waste-rock dumps in British Columbia (Canada) are presented
in Fig. 4. The majority of samples for three of the dumps would be classified as potentially acid
generating. The Island Copper dump is more complex, as potentially acid generating material
appears to 'outweigh' the non-acid generating material but may also occur in discrete zones
throughout the dump. This illustrates the need to integrate ARD testing with mine planning.
2.4.4 NP/AP ratio screening criteria
The issue of appropriate Neutralisation Potential/Acid Potential (NPIAP) ratios is a key area
of debate among the regulatory agencies and the mining industry. To date, there has not been
a comprehensive compilation of case histories of minesites with significant ABA data, NP/AP
ratios, and ARD problems. The use of NP/AP ratio is one of the most significant variables which
regulatory agencies are attempting to use as a prescriptive measure. A discussion of how POl
uses this variable as a screening criteria for classifying materials is warranted.
Ferguson and Morin (1991) and Cravotta et al. (1990) presented theoretical arguments suggesting
the NP/AP criterion separating potentially acid and non-acid generating samples could be about
2:1. However, in the database presented by Ferguson and Morin, no sample with an NP/AP
greater than 1 produced acidic leachate in 166 laboratory leaching tests. Moreover, there is
no clear documented evidence of rock with a NP/AP greater than 1 producing ARD under field
conditions.
11 MURRAY, ROBERTSON & FERGUSON
Defining the AMD Problem I.
300
2 250
a._
t-t-----t---+----1----1-----l--1-- ~
WESTMIN
~ Legend
(/!.
(')
0
I()
10
o-
()
8 200
Ol
2!:.
]j 150
c
rt-t- - 1 ~ - ~ t ~ : - ~ o 95%M'
LJMr:AsHI 5%AP
PIT I -
I - _ ACID -----+-=-.....-.___,
I I ISLAND COPPER I GENERATING Q)
0 100
NNP=O Line
a.
"'0
o
1 - i I -- --+--
' , I NON-AaD
GENERATING
oct 50
I
'1' --r-rl
MTWASH{UJIT -'
DUMP O 0 20 40 60 80 100 120
Neutralisation Potential (kg Caco
3
tt)
140 160
Fig. 4. ABA plot for acid generating waste rock dumps.
The NP/AP ratio may be considered a "safety factor" as used in other engineering analyses
where more definitive assessments of the actual risk of ARD are not available. Higher safety
factors are probably required for mines in wet climates where carbonate minerals may be
preferentially leached from the mine wastes relative to the oxidation of the contained sulfide
minerals. The criteria in Table 3 have been used by PDI as an initial screen of waste rock and
tailing ABA data.
Table 3. Categories used by PDI In screening ABA data.
Category NP/AP Range Description
Likely NP/AP < 1 Likely to generate ARD unless sulfide minerals
are relatively unreactive.
Possibly 1 < NP/AP < 2 Possible acid generating if neutralising
minerals preferentially depleted, coated or
unreactive.
Non-Acid Generating NP/AP > 2 Not expected to generate ARD.
In conjunction with the NP/AP ratios, the ABA data set for each reviewed operation is summarised
in a bar chart format. The objective is to illustrate the relative percentage of material which
falls within the various categories for the various material types such as ore, waste rock, and
tailing. An example application is shown in Fig. 5. The ore clearly has a higher probability of
generating ARD compared to the waste rock and tailing material. It is emphasised that this
form of analysis is another screening and planning tool and not the final assessment. This data
also needs to be evaluated with respect to the adequacy of the existing sample population.
MURRAY, ROBERTSON & FERGUSON
12
~
c:
0
:;
.0
,_
u;
i:5
.!
c.
E
111
en
100
90
80
70
60
50
40
30
20
10
0
Ore
Waste Rock
Sample Source
Tolling
Defining the AMD Problem I.
Category:
QNon-Acld
[SJ Possibly
II Likely
Fig. 5. ARD potential using NP/AP ratio screening criteria.
2.4.5 Modelling for spatial definition
New projects present a unique opportunity for the prediction of ARD. Historically, ARD prediction
for a new mine has involved the analysis of a couple dozen, often poorly selected, samples.
The principle adopted at PDI is that the overall economic significance to a project of the ARD
liability is comparative to the fundamental ore reserve inventory. A new mine would not be
engineered on the basis of only a few copper or gold assays, so why would a waste disposal
plan involving tens of millions of tonnes of waste be engineered on grossly inadequate ARD
data?
With the available resources, more complete ARD assessments are relatively easily achieved.
A phased approach is employed using all predictive tools, as discussed previously. If a potential
ARD problem is indicated, the first objective is to determine the order of magnitude and the
relative distribution of problematic material. The second strategy is to determine if there are
any key variables such as rock or mineral type which can be used to obtain a better view of
spatial variability in the ore body. This may require several thousand ABA analyses and extensive
computer manipulation of data, including the use of geostatistics and modelling.
This approach has been used to define waste rock which presents low, medium and high potential
for ARD. With this classification system, a mine plan can be developed which contains a range
of measures to prevent and to control ARD if it develops. Recent experience indicates that
specific waste types can be targeted in the mine plan for selective handling and disposal (Killam
and Robertson, 1994).
2.5 Research Programs
Over the past five years, PDI has supported the Mine Environment Neutral Drainage (MEND)
program and the British Columbia joint industry/government taskforce through involvement in
committees and projects. A total of exceeding $500,000 has been committed towards the overall
MEND budget of approximately $18 million. In addition to these research programs, PDI also
funds site-specific research at various minesites. Two of these research projects which have
yielded positive outcomes as discussed below.
13 MURRAY, ROBERTSON & FERGUSON
Defining the AMD Problem I.
2.5.1 Application of cover technology
Research is being conducted through the University of Saskatchewan to monitor and model
the moisture profile in soil covers on a number of waste rock dumps in different climatic regimes.
Compacted covers which maintain high moisture contents will significantly reduce the rate of
acid production by reducing oxygen diffusion and infiltration of precipitation. Results to date
have been encouraging with models validated in the field enabling optimum materials to be selected
for use as covers.
2.5.2 Subaqueous tailing disposal
The use of wet (saturated} covers on tailing has been studied over the past six years at several
sites in Canada. This MEND project has been managed by POl with an overall cost of
approximately $1 million. This work has successfully demonstrated that subaqueous disposal
effectively prevents sulfide oxidation and subsequent ARD generation in tailing deposits. The
results have been presented in papers at several conferences and work is ongoing to develop
design criteria for general application in the mining industry, although this may have limited
application in Australia.
3.0 CONCLUSIONS
POl has initiated an aggressive program to assess the ARD potential for all of its operations
and projects worldwide. To date, studies have been conducted at over 20 sites and a database
management system is under development to correlate all existing and future information. Research
programs are conducted and supported on an ongoing basis as an integral part of this program.
An important objective of the overall ARD program is to apply the predictive information to the
development of appropriate risk-based prevention strategies and/or controls on a site-specific
basis. This information allows the existing POl mines to modify their current waste management
practices where required. For mines which are in the planning process, the ARD assessment
program can be applied as a major variable in the evaluation of appropriate alternative waste-rock
and tailing disposal systems and practices.
All information and experience gained is directed toward yielding improved confidence in predicting
ARD potential, establishment of preventative measures, and development of precautionary strategies
across the corporation. The development of such complex programs is essential for building
confidence among the regulatory agencies for project permitting and approval. Complete programs
are essential for establishing credibility with all stakeholders and demonstrating ARD can be
managed effectively, despite some of the unknowns which will hopefully be resolved through
future research.
4.0 REFERENCES
Cravotta Ill, C.A., Brady, K.B.C., Smith, M.W., and Beam, R.L. (1 990). Effectiveness of the
Addition of Alkaline Materials at Surface Coal Mines in Preventing or Abating Acid Mine
Drainage: Part 1. Geochemical Considerations. Mining and Reclamation Conference
and Exhibition, April 23-26, 1990, Charleston, West Virginia.
Ferguson, K.O., and Morin, K.A. (1991 ). The Prediction of Acid Rock Drainage- Lessons from
the Database. Proceedings of the Second International Conference on the Abatement
of Acid Drainage, September 16-18, 1991, Montreal, Quebec.
MURRAY, ROBERTSON & FERGUSON
14
Defining the AMD Problem I.
Killam, R., and Robertson, J.D. (1994). Application of Modelling and Subaqueous Tailing Disposal
to the Mt. Milligan Mine. Proceedings of the Eighteenth Annual Mine Reclamation
Symposium, March 1994, Vernon, British Columbia.
Miller, S.D., Jeffrey, J.J., and Wong, J.W. (1 991 ). In-pit Identification and Management of Acid
forming Waste Rock at the Golden Cross Gold Mine in New Zealand. Proceedings of
the Second International Conference on the Abatement of Acidic Drainage, September
16-18, 1991, Montreal, Quebec.
15
MURRAY, ROBERTSON & FERGUSON
20
:._,t1l-UfHJ ACJd Uratnago Workshop (Eds N J Grundon and L C Bull). pp 21--31
SAMPLING AND WASTE BLOCK MODELLING
FOR ACID PRODUCING WASTE ROCK AT CADIA
G.M. JohnstonA, G. EastwoodA, P. Scott
8
, C. Joyce
8
and P. Ryan8
ANewcrest Mining Limited, 179 Great Eastern Highway, Belmont WA 6014.
8
AGC Woodward-Clyde Pty Ltd, 49 Park Road, Milton, Old 4064.
ABSTRACT
Over the years, predictive techniques for Acid Mine Drainage (AMD) and the natural processes
affecting acid generation have received considerable attention. Less emphasis has been placed
on the sampling strategy and the integration of testwork results with the geological data base.
The sampling strategy employed and the development of a geological model to determine the
spatial distribution and volume of acid producing material are often crucial for developing cost-
effective containment and mitigation measures. The geological model provides a link between
the occurrence, timing, destination and placement of mined waste rock. Careful and accurate
interrogation of a comprehensive geological database provides this linkage.
The AMD testwork program at Cadi a in NSW has been designed to identify the temporal selective
handling requirements that will be required during operation. Sampling involved collection of
representative combinations of all waste rock types identified within an extensive geological
database. These samples were subjected to static, kinetic and accelerated weathering predictive
tests. Preliminary results have been used to construct a waste block model for the deposit that
identifies the quantity and distribution of acid producing material. Integration of the block model
with the mine scheduling determines when the material wilt be mined.
1.0 INTRODUCTION
Newcrest Mining limited is currently undertaking a feasibility study for the Cadia copper-gold
deposit approximately 21 km south south-west of Orange in the Central Tablelands of New South
Wales (NSW). The Cadia project will consist of a large open-pit mine that will tre?t 15 million
tonnes of ore per year. The waste to ore stripping ratio is< 2:1. On-site processing facilities
will produce a copper concentrate with gold credits for smelting.
The project area is located in the Cadiangullong Creek Valley, a local creek that drains the southern
portion of the Mount Canobolas Park and Canobolas State Forest. The area is bounded on
three sides by moderately steep slopes rising to maximum elevation of 960 m towards the north.
Mean annual rainfall in the area is 900 mm spread fairly evenly throughout the year, and evaporation
is over 1300 mm year'.
1.1 Geological Setting
Cadia lies within the Lachlan Geosynclinal belt in south western NSW. The principal rocks exposed
in the region consist of Paleozoic volcanics, sediments and intrusives.
The Cadia copper-gold mineralisation is hosted in Upper Ordovician shoshonitic volcanics (the
Angullong Tuff). The Angullong Tuff at Cadi a contains lavas, tuffs, pyroclastics, limestone lenses,
and porphyritic and fine grained sub-volcanic rock types. Mineralisation at Cadia Hill is closely
associated with a porphyry-style hydrothermal alteration system that produced vein-controlled
JOHNSTON, EASTWOOD, SCOTT, JOYCE, & RYAN
21
'J"c,o<OII:j Li{ V<.iUid
mineralisation in the monzonite porphyry and immediate adjacent volcanic wall-rocks. The low-grade
gold-copper mineralisation that has been identified is hosted in an envelope of discontinuous
sheeted quartz veins that tend northwest and dip to the southwest
The sulfide minerals occur in two associations - as inclusions in quartz veins and as fine- to
medium-grained disseminations in the monzonite porphyry and volcanic host rocks.
1.1.1 Mineralogy
Mineralisation at Cadia Hill comprises rarely obseNed native gold, pyrite and several copper
sulfide species including chalcopyrite, bornite, digenite and primary chalcocite, and the iron oxides
magnetite and haematite.
The Cadia Hill mineralisation is indicative of a relatively oxidised and low total sulfide system,
as reflected in the generally low abundance of pyrite typically less than 1%, and the presence
of low sulfur copper minerals such as bornite and digenite.
2.0 SAMPLING STRATEGY AND TESTWORK
2.1 Sampling
The prediction of acid generation potential begins with an understanding of the geology and
mineralisation of the ore body and surrounding waste rock scheduled for removal. The general
approach adopted at Cadia is shown in Fig. 1. Preliminary screening testwork on selected
composited samples during the exploration phase indicated the potential for AMD. The results
of this screening program justified further testwork to assess acid producing potential. The
development of this program and the results to date are discussed in the following sections.
The geological database has been developed based upon the various geological categories
shown in Table 1. Geological logging was based upon 1 m inteNals, and all results were entered
into a central database.
The two major lithologic units at Cadia are volcanics generally located nearer the surface and
the monzonite intrusive. It was recognised that in each of these lithological units there was
a range of geochemical and physical conditions and types of mineralisation present Each of
these components will have different influences over the acid generating or neutralising capacity
of individual rock types (Ellison and Hutchison 1992). The type of sulfide minerals present and
the form in which sulfide mineralisation occurs (e.g. mineral particle size, whether disseminated
or vein, etc) will influence their reactivity. Other mineralogical factors affecting AMD behaviour
include:
(1) semi-conductor properties;
(2) crystal size;
(3) texture and morphology;
(4) degree of rock fracture; and
(5) spatial relationship with acid consuming materials in the host rock.
Representative samples of all combinations of rock types possible from the logging categories
were selected for testwork. In all some 36 separate categories of lithological, geochemical,
mineralogical and physical characteristics were identified.
JOHNSTON, EASTWOOD, SCOTT, JOYCE, & RYAN
22
EXPLORATION
PRE FEASIBILITY
ENVIRONMENTAL
IMPACT
ASSESSMENT
YES
PERMITTING
Initial Discussion with Project Geologist
Samples of ore and waste selected
In consultation with Project Geologist
Geographical , Geological, Palaeo-
environmental comparisons If available
Waste Block Modelling at Cadia.
Detailed discussions With Project Geologist
to define ore and waste rock units based ,,. ___ _,
on lithology, mlneralogy.fractures and
1
continuity. Inspect drill core
Fig. 1. Evaluating acid mine drainage potential at Cadla_
JOHNSTON, EASTWOOD, SCOTT, .JOYCE, & RYAN
23
Table 1. Sample selection criteria and description.
Hole Number (NC)
110
110
110
110
110
140
140
140
140
140
130
130
130
124
109
109
118
118
128
124
121
121
130
Interval
(m)
28-29
40-42
26-28
18-19
22-23
18-19
21-23
19-20
20-21
23-24
165-168
168-172
163-165
190-191
592-593
593-594
263-266
3-4
266-269
219-220
19-20
21-22
59-60
W;u;ta Rock Lithology % Pyl"IUI (loggild)
Monzonrte abundant
Monzonite modl!lral11
Monzonite -.bund;wt
Monzonrte abundant
Monzorute 01bundant
Monzonrte trace
Monzonite moder;J;te
Monzonite trace
Moru:omte .. ,.
Monzonite
Mon.zonrtf! abundant
Monzonite moderate
Motu:onite trace
Monzonite abundant
Monzonite
,.,.
Monzonite
""'"
Monzonite moderate
Monzonite
Monzonite trac6
Monzonite moderate
Volcanic moderate
Volcamc
"""'
Volcanic ilbundnt
WeatNtrlflgCiau
"(t<eyattlow)
sw
SN
MW
MW
FR
FR
FR
FR
FR
FR
sw
sw
sw
MW
HW
HW
FR
HW
FR
sw
sw
SN
FR
logg&i:J Oe:licrfptJon
No carbonate alteration
No carbonate alloro.tion
No carbonate alteration
No carbonate alteration
No Cilrbonate alteration
Nocarbonatealteratron
No carbonate alteratron
No c;ubonate alteration
No altoration
No carbonate alteration
Carhonate alteration
Carbonate alteration
Carbonate alteration
alteration
Carbonate alteration
Carbonate alteration
Garbonilte alteration with
calcite cavity fill
Carbonate altefOIIltonwith
calcite cavity fill
Cartxmate alteration With
calcitecavityfiU
No caruon.te alteration
No ca:rbonate alteration
Carbonilte With
l I I I I I I
130 60-61 abundant FR Carbonat!.! alt6ra!IOn with
52 36-37 Volcanic moderate
130 62-63 Volcanic trace
130 64-66 Volcanic trace
130 76-79 Volcanic abund01nt
130 23-24 Volcanic
130 18-19 Volcanic trace
130 19-20 Volcanic lrce
130 20-21 Volc01n1c traoo
52 29-30 Volcanic abund01nt
52 23-24 Volcanic moderate
99 27-29 Volcanic tl";;lce
99 29-32 Volcanic lrii.Ce
99 32-33 Volcanic
"""
99 41-42 Volcanic trce
120 325-327 Monzontte traoo
97 91-100 Volcanic abundant
100 149-157 Monzonite abundant
110 28-29 Monzonite abundant
110 26-28 Monzonite abundant
140 21-23 Monzonite modeflil.te
130 165-167 Monzonite 3bundant
124 190-191 Monzonrte abundant
JOHNSTON, EASTWOOD, SCOTT, JOYCE, & RYAN
24
FR
FR
FR
sw
sw
sw
sw
sw
sw
sw
&II
sw
sw
sw
FR
FR
FR
fNI
MW
FR
sw
-

Catbon01te alteril.tion with
calcite cavity fill
Carbonate alteration with
calcite cavity fill
Carbonate alteration with
calcite cavity fill
Cilrbonatc 01l!eration with
calcite cavity fill
with
calcite cavity fill
Carbonate alteration with
calcit6cav1tyfiH
Carbonate alteratiOn With
clcitecavityfi!t
C01tbonate alteratiOn with
calcitecvityfiU
Minor altertion
Minor carbonate alteration
Minor carbonate alteration
Minor carbonate Iteration
Minor carbonate 01lter;ation
Minor carbonate -.Iteration
Carbonate alteration with
cafcitecavityfiU
Carbonateo.lteration With
calcite cavity fill
Carbonate alteration with
caJcit6ca.Yityfill
No carbonate alteriltion
No carbonate o.Jteration
Nocarbonatealter.ttion
Carbonate alteration
Carbonate alteration
Waste Block Modelling at Cadia.
Using these categories, a search was made of the database to:
(1) confirm the presence of a rock type, i.e. were ail 36 possible rock types present
in the orebody;
(2) identify the interval containing the required rock characteristics; and
(3) determine how these rock types were distributed in the orebody.
A complete set of cross sections with the major rock characteristics were produced to show
the spatial distribution of characteristics across the deposit. Shown in Fig. 2 is a typical cross
section showing lithology, pyrite and carbonate content along with sulfide assays. On the basis
of this database, search samples of:
( 1) monzonite - high pyrite and low carbonate;
- high carbonate and low pyrite;
(2) volcanic - high pyrite and low carbonate; and
- high carbonate and low pyrite,
were selected for static, kinetic and accelerated weathering tests. Fifty six samples were selected
for static testwork and four samples for kinetic testwork.
120Ctn N
BOOm Rl
GEOCHEMISTRY LEGEND
"'''-\..Pyrite
l!imD COfb<XIatlil
ROCK TYPE
f:2Z3 Vo!ca()lc'
lliilllll1J Mon..zomre
WASTE GEOCHEMISTRY
SECTION 3820m E
Fig. 2. Typical geochemistry cross section.
Sampling was confined to waste rock outside the envelope of economic mineralisation. Tests
on tailings material were conducted separately and showed non-acid generating properties for
ali ore types except for high pyrite ores which recorded a low positive AMD potential. The pyrite
ores comprise less than 5% of the ore. Tailings material will be contained in a tailings storage
dam that will be capped and vegetated at the end of mine life.
JOHNSTON, EASTWOOD, SCOTT, JOYCE, & RYAN
25
Waste Block Modell1ng at Cadia,
2.2 Testwork Program
For each of the rock samples a complete geochemical analysis of major and minor elements
was performed. Some typical analysis results are shown in Table 2.
Table 2. Major and trace element geochemistry of waste rock at Cadia.
Sample As Cu Mn Se Zn Ca Fe Mg Tot. S
Number (mg kg-') (mg kg-') (mg kg-') (mg kg-') (mg kg-') (%) (%) (%) (%)
140 < 5 498 266 0.25 25 2.45 3.51 0.55 0.20
141 5 595 278 0.2 24 2.11 3.73 0.55 0.17
149 7 2600 199 1.06 57 0.46 3.07 0.61 0.03
150 6 2580 162 1.5 78 0.28 2.60 0.70 0.23
159 6 210 389 0.05 25 1.00 2.73 0.48 0.20
161 9 1090 213 0.65 18 1.02 3.62 0.45 0.19
164 20 690 157 0.7 15 0.21 2.81 0.68 1.38
2.2.1 Static testwork
The testwork carried out on 2-kg samples of the 58 samples consisted of:
(1)
(2)
(3)
saturation paste pH and EC;
acid/base accounting; and
leachate analysis.
2.2.2 Kinetic testwork
S-S so.-s
(%) (%)
0.19 0.01
0.16 0.01
0.06 0.01
0.12 0.02
0.19 0.01
0.16 0.02
1.29 0.01
Four 300 mm x 500 mm lysimeters were loaded with 10 kg of sample tor kinetic tests. Crushed
quartz was placed at the top and bottom to facilitate uniform flow. These tests have been running
tor 4 months under simulated average rainfall conditions. Humidity is maintained by humidifying
air entering the lysimeter. To simulate a fluctuating permanent water table in the dump, a side
tap at the base has been connected to a manometer that could be adjusted to various heights.
Another set of eight lysimeters have been constructed and will be used for accelerated weathering
tests involving repeated wetting and drying cycles based upon the average rainfall that occurs
on site.
3.0 RESULTS
The results of the static testwork are shown in Fig. 3.
The results of the static testwork show the distribution of material between acid producing and
non-acid producing ratios (maximum potential acidity/acid neutralising capacity). Fifty six percent
of material tested has a maximum potential acidity (MPA) value 2 3-fold the acid neutralising
capacity (ANC). A 'cut-off' ratio for the definition of potentially acid producing material of ANC:MPA
of s 3:1 and 0.2% S has been adopted based primarily upon the authors experience, and is
supported by Canadian AMD research experience (Brodie et al. 1991; Smith and Barton-Bridges
1991; Ferguson and Morris 1991; Ferguson and Robertson 1994).
JOHNSTON, EASTWOOD, SCOTI, JOYCE, & RYAN
26
t:
;f
::;:
Fig. 3.
Waste Block Modelling at Cad1a.
ANC (Kg H
2
S0
4
fT)
Static testwork on Cadla Hill waste rock.
The region between 1:1 and 3:1 ANC:MPA has been taken to be indeterminate because, in
this area, it is not possible to interpret the results as equivocally acid producing or consuming.
Nine percent of samples tested showed a high risk of potential acid producing material with
a net acidity exceeding 30 kg H,S0
4
t' waste rock.
Aside from the well documented limitations associated with static tests (they provide no information
on the rates of reaction), the graphical representation in Fig. 3 provides no information on quantities
of material, or the spatial occurrence of the material in the orebody. One way to overcome this
limitation is to use the results to construct a geological block model and this is discussed in
next section.
The kinetic testwork is designed to complement the static testwork and provide data on reaction
rates (i.e. information on quantity and changes with time). In particular, results that fall into
the indeterminate area on the static testwork graph can be resolved. Initial results from the
kinetic studies are shown in Figs. 4a and 4b.
The preliminary results indicate a slight increase in leachate pH from the columns with time,
and a decrease in leachate electrical conductivity. It is planned to run the kinetic tests for at
least 12 months to allow leachate water quality trends to emerged.
4.0 GEOLOGICAL BLOCK MODELLING
4.1 Background
A block model (as distinct from a geological model which provides spatial information on geological
units identified from drill hole information), calculates volumes (which can then be converted
to tonnages). A block model does this by defining and connecting adjacent cells with similar
properties. The spatial properties of these cells are defined by x,y, and z co-ordinates relative
to some reference point (Holding 1994).
Spatial controls tor determining the AMD category in which to place cells involved identifying
in the database common geological, physical or mineralogical characteristics that had been
shown to have the potential to produce AMD.
JOHNSTON, EASTWOOD, SCOTI, JOYCE, & RYAN
27
Waste Block Modelling at Cadia.
8.8 ,----------------------------,
8.4
8.0
7/
7.8
7.4.
7.2

NOV DEC JAN FEB
Month
1000r-----------------------------------------------------,
goo I - -1
BOO.\'::::: ...... =--------- "
7oo+


(i) 500 ..,
"'-
(.)
w
-+-
300
200
100 t------

NOV DEC JAN FEB
Month
Fig. 4.
Kinetic column testwork for pH (a) and electrical conductivity {b)
of leachate from Cadla Hill waste rock.
Once areas with similar AMD characteristics had been identified, the volume was calculated
by connecting cells with similar properties. With the volume of acid producing material defined
in space, it is possible to integrate these volumes into the mining schedule to determine how
much, and when, potential AMD material will be produced during mining.
JOHNSTON, EASTWOOD, SCOTT, JOYCE, & RYAN
28
Waste Block Modelling at Cadia
4.2 Results
The block model was run for potential acid producing material identified through the static testwork.
Acid producing material was defined to included material that had been shown to have an ANC:MPA
ratio< 1, as well as material with an ANC:MPA between 1 and 3. A summary of the parameters
used for block modelling is shown in Table 3. The model was run at a 0.5 g t' equivalent gold
cut-off, and the results are shown in Table 4.
Table 3.
Parameter
Summary of block modelling
parameters.
Interpolation Nearest Neighbour
technique
Search ellipse 600 m radius
Cell size
4:1 anisotropy i.e. 600 m
along strike and down dip
and 150 m perpendicular
to strike
2m North 12.5 m East
10mR.L.
Table 4. Results of block modelling for
waste AMD at a 0.5 g t'
equivalent cut-off.
Category
Potential AMD
Indeterminate
AMD
Total
Percent of total waste
volume
22%
14%
36%
5.0 DISCUSSION AND CONCLUSIONS
A typical cross section showing the spatial
characteristics of interpolated static test
results using block modelling is shown in
Fig. 5. The results of the kinetic testwork
will be incorporated into the model when
they become available. A lowering of the
economic cut-off grade below 0.5 g !"
1
will
significantly reduce the percentage of acid
producing waste because the envelope of
economic mineralisation will expend greatly
from that shown in Fig. 5 and Table 4. In
addition, geological hard boundaries for
some sections of the ore body are still
being developed and are expected to
reduce the total volume of acid producing
material.
Validation of the model is currently being
undertaken. This will include duplicate
testwork and random sampling to detemnine
the predictive power of the model. The
next phase of work will be to couple the
waste block model results with the mining
plan to construct the mine waste schedule.
The mine waste schedule will be used to
develop a waste handling program and a
waste dump design to control potential
AMD production.
Obtaining reliable samples for acid prediction testwork is a difficult process and is often not achieved
due to a lack of understanding of block modelling techniques, and a lack of integration between
geochemistry and applied geology.
In order for the acid-generating properties of rock types within an orebody to be representative,
adequate sampling of material is essential {Rymer et al. 1991 ). Probability theory provides a
readily available means for determining a suitable sample size for a given set of geochemical
groups within each lithic unit. It is apparent that the number of samples taken to evaluate a
potential mine will often not be adequate to satisfy sampling probabilities, to achieve the level
of accuracy needed for accurate AMD prediction. Economic consideration will often preclude
the drilling of more holes to provide the necessary sample density or testing of large numbers
of samples for AMD characteristics
JOHNSTON, EASTWOOD, SCOTT, JOYCE, & RYAN
29
Waste Block Modelling at Cadia.
tlli1il
n
-
NON-A.C!O PROOUCING
INTERl.4EOIATE
AGIO PROOUCING
CADIA PROJECT
ANC/MPA BLOCK MODEL
SECTION 3620E
' """'
. '
Fig. 5.
Cross section of block modelling results at Cadia.
The design of an AMD program must therefore balance the conflicting requirements of generating
enough useful data that can be used to make predictions and containing costs against the potential
costs of remediation of unanticipated or incorrectly diagnosed AMD problems.
It is important to ensure that enough relevant geological information is recorded during the
geological logging of the core so that it is possible to tailor the sampling program to take account
of the geological features of the deposit. By utilising all the available information it is possible
to design a program that meets the above objectives and is cost effective. It the geology is
not considered in sufficient detail ear1y on or the geological database does not contain the necessary
information then the ability of any program to correctly predict AMD properties is severely
compromised.
The use of sophisticated computer block modelling techniques allows the results of an AMD
program to be used to identify the spatial distribution and volume of acid producing material
within the deposit. This information is a prerequisite if cost-effective waste handling programs
are to be developed that can be integrated with the mine development schedule for the deposit.
In summary, linking the sampling to the geology and block modelling provides the following
advantages:
(1) a cost effective approach;
(2) the total volume of problematical material can be determined; and
(3) the timing of when this material becomes available can be determined.
The Cadia example documented above illustrates how to make the most use out ol the geological
database when selecting samples for an AMD program and how to use the information provided
from the testwork program to construct a block model to assist in management strategies.
JOHNSTON, EASTWOOD, SCOTT, JOYCE, & RYAN
30
Waste Block Modelling at Cadia.
6.0 REFERENCES
Brodie, M.J., Broughton, L.M., and Mac Robertson, A. (1991 ). A conceptual rock classification
system tor waste management and a laboratory method for ARD prediction from rock
piles. Proceedings: Second International Conference on the Abatement of Acidic Drainage,
Sept 16 to 18, 1991, Montreal Quebec.
Ellison, RD., and Hutchison, I.P.G. (1992). Mine Waste Management. (Lewis Publishers: Ann
Arbor.)
Ferguson, K.D., and Morin, K.A. (1991 ). The prediction of acid rock drainage- Lessons from
the database. Proceedings: Second International Conference on the Abatement of Acidic
Drainage, Sept 16 to 18, 1991, Montreal Quebec.
Ferguson, K.D., and Robertson, J.D. (1994). Assessing the risk of ARD. Proceedings: International
Land Reclamation and Mine Drainage Conference and Third International Conference
on the Abatement of Acidic Drainage, April 24 to 29, 1994, Pittsburg PA.
Holding, S.W. (1994). 30 Geoscience Modelling: Computer Techniques tor Geological
Characterisation. (Springer-Verlag: New York.)
Rymer, T.E., Skousen, J.G., Renton, J.J., Zondlo, A.M., and Ziemkiewicz, P.F. (1991 ). Undetected
sources of detectable acid mine drainage: Some statistical considerations in rock sampling.
Proceedings: Second International Conference on the Abatement of Acidic Drainage,
Sept 16 to 18, 1991, Montreal Quebec.
Smith, A., and Barton-Bridges, J.B. (1991 ). Some considerations in the prediction and control
of acid mine drainage impact on groundwater from mining in North America. Proceedings
of the EPPIC Water Symposium, May 16-17, Johannesburg, South Africa.
,JOHNSTON, E:ASTWOOD, SCOTT, JOYCE, & RYAN
31
32
Second Australian Acld M1ne O r a ~ n a g e Workshop (Eds N.J Grundon and L C. Bell). pp 33--42.
ADVANCES IN THE PREDICTION OF ACID
GENERATING MINE WASTE MATERIALS
S. Miller and J. Jeffery
Environmental Geochemistry International Ply. Ltd.,
20 Rowntree Street, Balrnain, NSW 2041.
ABSTRACT
Predicting the acid generating nature of individual rock types and mine waste is the fundamental
first step in any mine waste geochemistry assessment and provides the foundation for developing
an acid rock drainage and acid spoil control and management plant. Feasible options for
management are determined mainly by the geochemical nature of individual rock types, the
quantity of each geochemical rock type and the mining schedule.
This paper presents an overview of the current prediction technologies and recent developments.
Both static and kinetic tests are described and common interpretation errors and pitfalls with
these procedures are discussed. Recent development in prediction and field identification of
potentially acid forming materials are presented including a discussion of the net acid generation
(NAG) test. Research into the NAG test applications is being carried out under an AMIRA research
project. This is an important development and provides the industry with a useful tool for identifying
and managing potentially acid generating mine waste materials.
1.0 INTRODUCTION
There are two main purposes for acid generation prediction work associated with mine wastes.
One is to predict the behaviour of individual rock types within the deposit so that a management
plant can be developed to control acid drainage and prevent acid toxicity within the revegetation
layer. The second is to predict the overall behaviour of the waste rock dump, process waste
emplacement, open pit or underground void and is an extremely complex and difficult task.
Predicting the field behaviour of a constructed dump is influenced by many factors including
the adopted management strategy which may involve deep burial of potentially acid-forming
material, blending with acid-consuming materials, placement of constructed layers or zones
to control oxygen ingress etc. The physical and hydrological properties of the material and
prevailing climatic characteristics significantly influence the field behaviour and local impact.
This paper focuses on predicting the acid generating nature of rock types. This is the fundamental
first step in any mine waste rock geochemistry assessment and provides the foundation for
developing an acid rock drainage and acid spoil control and management plan. Feasible options
for management are determined mainly by the geochemical nature of individual rock types, the
quantity of each geochemical rock type and the mining schedule.
The level of prediction and control required will depend on the environmental setting and sensitivity.
In drier climates where leachates may not be a concern, the objective may be simply to ensure
that potentially acid-forming materials are not placed in or near the plant rooting zone whereas,
where leachates are a concern, control of acid drainage may be important. In all cases it is
necessary to conduct basic investigation to categorise the geochemical rock types. The extent
of the investigation required will vary from a simple evaluation of selected rock types to a detailed
33
MILLER & JEFFERY
"'. - ~ " ' - ' ' " '
program which can be integrated with geological and mining block modelling.
A key component of an ARD/AMD prediction investigation is sampling. It is essential that all
rock types and waste types are included and that the number of samples is sufficient to provide
at least the boundaries of variability. The samples should be selected to allow development
of possible correlations with geological rock types or stratigraphic position and allow geochemical
rock type mining schedules to be prepared.
This paper presents an oveNiew of the current prediction technologies and recent developments.
Both static and kinetic tests are described, and common pitfalls with these procedures are
discussed. Recent developments in prediction and field identification of potentially acid-forming
materials are presented including a discussion of the net acid generation (NAG) test. Research
into the NAG test applications is being carried out under an AM IRA research project sponsored
by seven Australian mining companies. This is an important development and provides the
industry with a useful tool in avoiding and managing potentially acid-generating mine waste
materials.
2.0 PREDICTION OF ACID GENERATION POTENTIAL
Prediction techniques can basically be categorised into three groups:
(1) Geological assessment;
(2) Geochemical static tests; and
(3) Geochemical kinetic tests.
2.1 Geologic Assessment
Geologic comparison with other mine deposits and mine sites is a commonly used technique.
It is based on the assumption that mine sites having the same or similar paleoenvironmental
or geologic characteristics should have the same potential for acid generation.
Paleoenvironmental models have been used to predict the quality of coal mine drainage on
a regional scale. Acid drainage can be linked to the paleoenvironment of coal formation. Pyritic
sulfur formed in marine and brackish water environments generally has a higher capacity in
acid generation than that formed in fresh water environments. Fine-grained 'framboidal' pyrite
is commonly formed in marine environment and is more reactive than many other forms of pyrite.
Geological models are very useful for predicting the potential for acid generation. Some field
geologists are now logging and recording information relevant to acid drainage prediction (this
needs to be encouraged in all exploration drilling program). The following factors, which are
partly due to the geological environment influence generation of acid drainage:
(1) Oxidation state of the minerals;
(2) Sulfide mineral content;
(3) Sulfide mineralogy;
(4) Carbonate mineral content; and
(5) Rock alteration and hardness.
These factors provide the necessary information for the initial step in classifying mineral deposits
in terms of their potential to produce acid drainage.
MILLER & JEFFERY
34
Prediction of Acid Generating Materials.
Mineralogical examination of mine rock samples can provide very good insight into the potential
for ARD and is highly recommended for all mine sites. The basic data required includes microscopic
examination (reflected and transmitted light) and XRD analyses of carbonate species.
2.2 Geochemical Static Tests
A static test is a simple, rapid and relatively inexpensive test to qualitatively predict the acid
forming potential of a mine waste material. Static tests evaluate the balance between acid
generation processes (oxidation of sulfide materials) and acid neutralising processes (dissolution
of alkaline carbonates, displacement of exchangeable bases, and weathering of silicates).
The two approaches to static test are the acid-base account procedure and the hydrogen peroxide
direct oxidation procedure. There are a number of variations with the acid-base approach but
generally, in the Asia Pacific region, the net acid producing potential (NAPP) method is used.
The hydrogen peroxide-based net acid generation (NAG) test is now becoming more commonly
used, not just as a confirmation test and field test, but its application for predicting reactivity
and the lag period are very encouraging. Both these procedures are discussed below.
2.2.1 Net acid producing potential (NAPP)
The Net Acid Producing Potential ( NAPP) is calculated simply by subtracting the acid neutralising
capacity (ANC) from the calculated maximum potential acidity (MPA) using the following equation:
NAPP (kg H
2
S0
4
r') - MPA- ANC (kg H
2
S0
4
t-')
[1]
where MPA is (%S x 30.6).
The NAPP can be calculated using the total sulfur, total sulfide sulfur or pyritic sulfur content.
Each refinement results in a better estimate of the MAP but also increases the analy1ical costs.
The stoichiometric conversion factor of 30.6 assumes all sulfur is present as reactive pyrite and
that the reaction proceeds to completion. An adjustment to this factor is required if other acid
generating sulfides such a pyrrhotite or chalcopyrite occur. Allowance may also need to be
made for non-ferrous sulfides such as galena or sphalerite which are essentially non-acid
generating. Sulfates and organic sulfur are also not acid generating and must be considered
in some situations.
Generally, non-coal deposits contain little or no organic sulfur but can contain substantial sulfate
sulfur. On the other hand, much of the sulfur in coal deposits may occur as organic sulfur and
thus the total sulfur content will significantly over-estimate the acid potential.
When reporting the NAPP it is important to indicate the type of sulfur used in the calculation.
Hence, an understanding of the sulfur forms in the mine waste is essential for the correct use
of NAPP procedure. When dealing with coal reject and tailings materials, it is recommended
tha pyritic sulfur always be determined. For other types of mine rocks and wastes, selected
samples should be analysed tor the forms of sulfur to confirm the suitability of using total sulfur
in the acid-base calculation. This approach is recommended to minimise the risk of over-estimating
the acid-forming potential of the material.
The Acid Neutral ising Capacity (ANC) measures the ability of a sample to neutralise acid generated
from sulfide oxidation. The ANC is generally determined by reacting a sample with a known
amount of standardised HCI or H,SO, and back titrating to determine the amount of acid consumed.
35
MILLER & JEFFERY
Prediction of Acid Generating Materials.
Some problems encountered with this procedure occur if the ANC is relatively high and the amount
of acid added is not sufficient to react with all carbonates. It is advisable to check the pH before
back titrating to ensure that sufficient excess acid has been added.
At some operations, alternative procedures are used such as determining Ca, Mg and Mn on
the HCI digest and calculating the ANC by assuming these elements are all present as carbonates.
To provide a better understanding of the availability of the ANC, a slow titration method is
recommended. The sample is slowly titrated with a set amount of acid of known strength and
a titration curve generated (Fig. 1 ). The buffering characteristic of the material can then be
evaluated. The plot shown in Fig. 1 is for a sample with a total ANC of 55 kg H
2
S0
4
t' However,
the titration curve indicates that, after adding 35 kg H
2
S0
4
r', the pH begins to fall rapidly and
indicates that not all the ANC will be available for neutralising sulfide generated acidity.
9.00
8.00
7.00
6.00
pH
5.00
4.00
3.00
2.00
i.
I...._
--
0
~
--------
10
.....
......
.......
' i\_
'
"'
\
--- --- - - -- --- -------
20 30 40
Acid Added (kg H2S04/I)
Fig. 1. Characteristic acid buffering curve (ANC).
- --- -- --
50
The titration procedure may provide a more accurate estimation of the neutralising capacity
in the waste since no heating is involved, and the pH of the test is more realistic than the very
low pH of the back titration procedure. This is an important consideration if dolomites or ferroan
dolomites occur which have a low solubility at normal acid drainage pH (3.0-3.5). However,
if there is a close association between the sulfides and carbonates in the material, the pH at
the reaction zone may be significantly lower than the bulk leachate making even relatively insoluble
carbonates reactive under certain conditions.
It is strongly recommended that representative samples from a site are evaluated by this procedure,
particularly if a significant amount of the waste has a border line NAPP result.
Theoretically, a sample with NAPP > 0 has a MPA greater than the inherent ANC and therefore
MILLER & JEFFERY
36
Predict1on of Acid Generating Materials.
has a potential to generate acid whereas a sample with NAPP s; 0 has an ANC greater than
or equal to the MPA and has the capacity to neutralise any acid generated by the contained
sulfides.
The NAPP procedure is not a definitive test, as indicated above, and needs to be supported
by kinetic test work and NAG test to confirm that a sample has the potential to generate acid
and to provide an indication of the lag period or exposure time necessary before acid conditions
become established. These aspects are critical for developing management strategies and
assessing the environmental significance of a particular deposit or waste material.
However, the standard NAPP procedure does allow a large number of samples to be tested
and screened into relevant geochemical groups with respect to the theoretical capacity to generate
acid. When used in this way, as a first pass screening test, the limitations and pitfalls of the
procedure are not a major concern. It is strongly recommended that the NAPP procedure is
not used as a definitive test on which planning and operational decisions are based.
2.2.3 Net acid generation test (NAG)
This is a test which directly evaluates the net acid generation potential without measuring the
acid generation potential and the acid neutralisation capacity separately. Environmental
Geochemistry International Pty ltd (EGi) have been commissioned and funded by members
of the Australian Mineral Industries Research Association (AM IRA) ltd to further develop the
NAG test as a simple low cost procedure for the identification of acid forming mine rock and
process residues that could be utilised for environmental management and planning of waste
disposal. The companies sponsoring this research are Billiton Australia, BHP Australia Coal
Pty Limited, Carpentaria Gold Pty Limited, Hamersley Iron Pty Limited, Peak Gold Mines Pty
Limited, Placer Pacific Limited, P .T. Kelian Gold Equatorial Mining, and Western Mining Corporation
Limited - Kambalda Nickel Operations.
The overall project involves direct comparison of NAG, NAPP and leach column results for an
extensive and wide ranging group of mine rock and process residue samples supplied by the
project sponsor companies.
The aims of this study are:
( 1) To develop the NAG procedure as a simple low-cost test for identifying acid-forming
mine rock and waste materials for use in pre-mining and operational phases of
mining ventures;
(2) To study the kinetics of the NAG reaction by measuring pH and temperature profiles
during sample reaction and changes to these profiles under specific conditions;
(3) To evaluate whether the results can be extrapolated to the field situation for prediction
of the kinetics of the acid generation and neutralisation reactions in the field and
the exposure time necessary before acid formation occurs (i.e., the lag period);
and
(4) To assess on-site applications of the NAG test, including in-pit monitoring of acid-
forming materials.
The project began in October 1993 and is due to be completed in May 1995.
37
MILLER & JEFFERY
Prediction of Acid Generating Materials.
The Net Acid Generation Test (NAG) involves the addition of unstabilised hydrogen peroxide
to a sample of waste rock or tailings then measurement of the solution pH (NAGpH) and acidity
after the reaction with hydrogen peroxide is complete. The hydrogen peroxide reacts with sulfide
minerals, generating heat and acid. The acid in turn reacts with any inherent acid neutralising
minerals in the samples. Therefore, the amount of acid remaining in solution at the completion
of the reaction, which is quantified by titration, represents the net amount of acid generated
by the sample. The NAG test requires minimal laboratory equipment to carry out and a standard
NAG test can be completed within hours. This means the NAG test is a convenient and appropriate
for field test for operational use.
The standard NAG test is now generally accepted as a tool for assessing the acid potential
of a mine waste rock and process tailings. Problems can be encountered with organic containing
waste but these can be overcome by field calibration and modification of the procedure to ensure
that all organic residues are decomposed prior to measure NAGpH.
Interpretation of the standard NAG test requires some site input and calibration. However, generally
a NAGpH < 4 indicates that the sample is potentially acid forming where as a NAGpH ~ 4 indicates
that the sample is non-acid forming. Standard additions of NaOH or full titrations can then be
used to identify low capacity PAF material from higher capacity PAF material for refining waste
management operations.
A major objective of the AMIRA project is to evaluate if the kinetics of the NAG test reactions
can be related to real time kinetics of sulfide oxidation and acid generation in the field. Although
the results are still being evaluated and a longer time period will be required to confirm the
behaviour of some samples, the preliminary findings indicate that the NAG test can be used
to categorise materials on the basis on their lag period.
Prediction of the lag period is an important consideration for waste management as it will detennine
if materials need to be treated or buried immediately, can be left exposed for a period of time
before covering or, where sulfide reactivity is very low, only minimal controls may be required.
Figure 2 shows the pH and temperature NAG reaction kinetics for a sample which is classified
as PAF (high capacity) but has a relatively long lag period. Based on the information available
to date, a sample with a monitored NAG profile similar to Fig. 2 would be expected to have
field lag period in excess of 5 years. Other NAG profiles indicate difterent reactivity and lag
periods and full details of the research findings will be reported at completion of the current
AMIRA test work.
2.3 Geochemical Kinetic Tests
Commonly used kinetic tests include humidity cells, column tests, batch tests, soxhlet reactor
tests, field lysimeters, trial dumps and barrel test. The NAG test, discussed above under static
tests, can also be considered a kinetic test as the data generated will provide information on
reaction kinetics.
The major objectives of kinetics tests are:
(1) To provide real time data on the kinetics and rate of acid generation and acid
neutralising reactions under laboratory controlled or on-site conditions;
(2) To provide information on metal release and drainage/seepage quality; and
MILLER & JEFFERY
38
6
5
t=
4
p-l
3
2
~
0
0
-
f-- Temper ture
---&
r- d1
. . . . . .
~
~
..
~
!'q
\
.I
........

100 200 300
Time
(minutes)
Fig. 2. Monitored NAG test plot.
Prediction of Acid Generating Materials.
'
ho.
~ _pa
400
80
60
40
l
20
I
0
500
Q)
..
. B ~
cdC)
...
Q) bll
o..,
s ~
Q)
8
(3) To evaluate various treatment options such as covers, liming, layering, inundation
and chemical addition (bactericides etc.).
The methods attempt to simulate the weathering process that leads to acid and base dissolution
under lalooratory controlled or site conditions. The kinetic test results provide information on
the acid-generation characteristics and indicate if the rate of acid generation is significant or
negligible, over what period it may occur, and hence, over what time period controls would be
required. This is important in the planning of the disposal strategy, cover design and final
decommissioning.
It is common to most of the kinetic tests that water is added to a sample, the mixture is allowed
to incubate for a certain specified period, acid producing and consuming reactions are allowed
to proceed, and samples of leachate or extracts are collected and analysed. There are various
methods of adding water to the reaction flasks or columns depending on the objective of the
tests selected. Water application includes continuous or intermittent inundation, continuous
or intermittent simulating rainfall, and alternate exposure to humid atmospheres and leaching.
Sealed barrel test are used to determine oxygen consumption rates for dump oxidation
assessments.
Since kinetic methods provide information on the time of overall acid generation, the tests are
required to continue tor months or years, and generally the number of samples tested is limited.
39
MILLER & JEFFERY
Prediction of Acid Generating Materials.
The major parameters to be monitored in kinetic tests are the trends in pH, sulfate, acidity or
alkalinity, and metals. The pH of the collected leachate helps to identify the stage of the acid
drainage process. Sulfate production can be related to other rate of sulfide oxidation. The acidity
or alkalinity gives an indication of the rate of acid production and acid neutralisation. Metal analyses
are important to evaluate metals solubility and leaching behaviour.
The most commonly used kinetic test are humidity cells and column leach test. A major problem
with the humidity cells is that a high solution to solids ratio is used and the concentration of
constituents in the extracts is very low. This can effect geochemical processes within the sample
and also provides very little data on leachate quality. In addition, the humidity cell test was
originally developed for coal overburden where the lag period is generally short and the 'standard'
test is run for only 10 weeks. For non-coal materials this time period is inadequate and can
give false results.
Column test also have drawbacks and a frequent problem is that the sample becomes water
saturated or partially saturated, and interpretation of the results is therefore complicated. Provided
this problem is addressed, column test are a better representation of the field situation that the
humidity cells and provide more appropriate data on the geochemical process and leachate
chemistry. Column test also allow treatments to be tested and compared which is a major
advantage over humidity cells.
Figure 3 shows the standard column set up used by EGi. The columns are placed under heat
lamps so that the sample dries between water additions to ensure an adequate air void content
throughout the column.
Approximately 2.5 kg of samples should be placed in the column. Waste rock should be crushed
to nominal 4 mm size and tailings used in the as received state. The sample is placed in the
column and initially leached with deionised water at a rate of 400 ml kg-'; when drained, the
leachate is collected and labelled week 0. Leaching is then carried out at a rate of 100 ml
kg' for three weeks and 400 mL kg' on week 4. The leachate is collected after the week 4
addition. For the initial wetting-up and week 0 leachate, the water can be added over a 1 to
2 day period, depending on the permeability of the sample. All other water additions should
be added on one day (the same day) each week. Between water additions, the column should
be exposed to drying conditions. The residual water content in the column can be checked
by determining the initial column plus sample set up weight and comparing this to the total weight
prior to water additions. The approximate air void content can then be calculated.
Where a heat lamp is used, the distance between the lamp and the sample is adjusted to maintain
the surface temperature at about 35 C.
The objectives of kinetic test need to be clearly defined before deciding what type of test is
appropriate. With respect to evaluating the kinetics of sulfide oxidation and acid generation
under oxidising conditions, it is recommended that a standard procedure be adopted by industry.
The reason for standardising the procedure is to enable different rock and waste types to be
compared thus providing better prediction in the future.
3. CONCLUSION
The overall objective of this paper is to provide operators, government and authorities and others
involved in ARD/AMD management with fundamental information on prediction techniques for
defining waste rock geochemical types for ARD/AMD management on site.
MILLER & JEFFERY
40
Prediction of Acid Generating Materials.
SHELF
~ H e a t larrp
175mm
1 Li tro I LEACHATE
COLLECTION
VESSEL
BENCH
"'
8
~
Fig. 3. Column set up for evaluating acid generation and acid
neutralisation process.
The paper condudes that the simple acid base procedure (NAPP) is appropriate for initial screening
of samples. The net acid generation (NAG) test has proven to be reliable for evaluating the
acid-generating nature of samples, and the results of these two tests provide a sound basis
for classifying rock types for operational management.
The results of these tests allow rocks to be classified into the following categories:
(1) Acid consuming (ACM);
(2) Non-acid forming (NAF);
(3) Potentially acid forming - low capacity (PAF-LC); and
(4) Potentially acid forming -high capacity (PAF-HC).
Where further information Is required on the kinetics of the acid generation and acid neutralising
processes, column tests are recommended. The need to standardise the column test set up
and procedure is highlighted and a proposed method presented.
The NAG test has been proven as a very useful tool in the prediction process and for operational
monitoring. The test is also being developed for assessing the lag period and reactivity of a
sample. The results of this research will be completed by mid-1995.
41
MILLER & JEFFERY
42
Second Australian Acid M1no Dra1nage Workshop (Eds. N.J. Grundon and L C. Bell), pp 43-SL
A MANAGER'S OPERATIONAL PERSPECTIVE OF
AMD MANAGEMENT AT THE PAJINGO GOLD MINE
I.A. Tredinnick and P.J.Cornwell
Battle Mountain Australia Inc, Pajingo Gold Mine, P.O. Box 237, Charters Towers, Old 4820.
"The old gold mining industry usually paid little attention to the environme/11. Victorians
in the 1850's could tell when a new digging had opened up forty miles upstream:
the river water downstream quickly changed colour with the clays and gravels that
had been overturned upstream."
"Stawell, which in the late 1870's was the deepest goldfield in Australia, announced
its presence to the approaching traveller by the taste of sulphur from the kilns where
the gold bearing pyrite was roasted. People did not see Stowell as they approached:
they tasted it."
"Gold mining is not only seen by the general public as less productive of wealth than
a century ago. It is also seen by an influential minority, as more destructive of the
environment than in the past. This is a curious change because there is no doubt that
most gold producers in Australia are far more sensitive to their environme/11 than they
were a century ago." (Blainey 1991).
ABSTRACT
The Pajingo Gold Mine is situated 53 km south-east of Charters Towers, North Queensland.
The mine is owned and operated by Battle Mountain (Australia) Inc., a wholly owned subsidiary
of Battle Mountain Gold Company of the U.S.A.
Pajingo was a virgin discovery in late 1984, and mining commenced in August 1987. To date
some 350,000 oz of gold and 1 million oz of silver have been produced.
One of the many planning considerations in any mining operation is environmental management.
Pajingo has adopted a proactive role in identifying the problems, planning and implementing
a strategy, and monitoring the results. The key environmental issues at Pajingo include preventing
acid mine drainage, minimising the areas of disturbance, identifying suitable rehabilitated land
forms and adopting a tails dam capping strategy.
As an operator the measure of success is a practical, efficient, cost effective solution. It is essential
for operators in the industry to integrate rehabilitation planning and costing into any project at
the feasibility and development stage. Specific practical solutions need to be developed for
each identifiable problem, and continua/liaison between Consultants, Regulators, Contractors
and Landowners must occur to produce an acceptable result. The future of the industry depends
on it.
1.0 INTRODUCTION
Acid mine drainage (AMD) was identified as being a potentially significant environmental issue
at the Pajingo Mine early in its history. A combination of corporate perspective, site management
philosophy and operational practices has lead to a series of AMD management strategies. The
following text overviews the potential magnitude of issues and provides details of the management
issues specific to the Pajingo Operation.
43
TREDINNICK & CORNWELL
AMD Management at the Pajingo Gold Mine.
2.0 THE PAJINGO MINE AND BATTLE MOUNTAIN
The Pajingo Mine site is located 75 km by road south-east of Charters Towers and some 150
km south-southwest of Townsville in North Queensland. The mine is operated by Battle Mountain
(Australia) Inc. (BMAI), a wholly owned subsidiary of the Battle Mountain Gold Company (BMGC)
of the U.S.A. In addition to the Pajingo Mine, BMGC has attributable production from live other
mines in North America, South America and Australia. The total attributable gold production
for BMGC during 1994 was 486,000 oz of gold, of which 30,000 oz came from Pajingo.
The Pajingo Mine was established upon the grassroots discovery of the Scott Lode orebody
in 1984. Mining operations commenced in September 1987 and reached full production in
December 1987. Initial reserves indicated a relatively short mine life with reserves being depleted
after6 years (1993). The Cindy vein was discovered in 1991 and mining commenced in February
1994. The Scott Lode orebody was mined out in 1993 and production from the Cindy vein is
due to be completed in May 1996. Based upon project to date production and current reserves,
total production will be approximately 407,000 oz of gold. Table 1 is a summary of the mined
areas, mining period, and estimated gold production.
Table 1. Production summary
Ore body Date Production Head grade
(oz) (g Aut-')
Scott Lode open cut Sept. 87 - Aug. 92 348,000 9
Scott Lode underground Dec. 92 - Dec. 93 9,000 7
Cindy open cut Feb. 94 - June 94 10,000 6
Cindy underground Aug. 94 - (May 96) 40,000 7
Milling completed (June 96)
Ore is treated in a conventional carbon-in-leach/carbon-in-pulp treatment plant with an annual
throughput of 180,000 t year-'. Up to 1993 the average head grade was 10 g Au c'. The head
grade thereafter will/has ranged between 5 g Au c' and 7 g Au c'.
During the course of the project, BMAI has employed between 35 and 40 permanent employees
in the mining, treatment, supply and administration departments. All mining activities (open
cut and underground) have been carried out by contractors.
3.0 ENVIRONMENTAL MANAGEMENT
3.1 Corporate Perspective
BMGC has a formal statement of environmental policy and principles. This is detailed under
the following sections: Management Commitment, Employee Participation, Conservation of Natural
Resources, Emissions to Air Water and Soil, International Operations, Leadership, Public Policy
and Assurance By Senior Management.
"BMGC's international operations are designed to meet environmental regulatory requirements
TREDINNlCK & CORNWELL
44
AMD Management at the Pajingo Gold Mine.
and to achieve the same long term expectations as our U.S. operations with respect to
Environmental Policy and Principals." (BMGC 1993). BMGC recognises the advanced environmental
requirements of Australian legislation and accepts these requirements as suitable operational
conditions. This is in contrast to BMGC operations in parts of South America where low
environmental standards have not been acceptable to BMGC. In these cases BMGC imposes
a combination of U.S. and local environmental standards upon its operations.
Pajingo management via the Vice President, Operations is required to provide an annual
environmental update to the Environmental Affairs and Ethics Committee of the Board of Directors
of BMGC.
3.2 Site Philosophy
Acid Mine Drainage is viewed on site as being potentially the longest lived legacy of the Pajingo
Operation. This, combined with corporate philosophy, a desire to be proactive, the short mine
life, and a small work force has shaped environmental and, in particular, AMD management
practices on site. The aim is to maximise the available favourable materials and develop a self
managing AMD strategy.
The short mine life meant that AMD management had to be dealt with up front. There were
not sufficient reserves to allow extensive field trials to be conducted and the results used to
develop AMD management plans. The object was to develop an AMD management program
that provided cost-effective, up-front strategies, avoided extensive post-mine management
commitments, and provided sufficient up-front data to allow realistic and appropriate environmental
standards to be agreed with Regulators. The relatively small production base meant that research,
AMD management solutions and on-going monitoring must be kept in proportion ($)to the rest
of the operation.
4.0 SCALE/MAGNITUDE OF POTENTIAL SOURCES OF AMD
4.1 Quantities
The high waste to ore ratios for the Scott Lode open pit (14:1) and the Cindy open pit (22:1)
have resulted in relatively large quantities of waste being generated when compared to the quantity
of ore processed. Tables 2 and 3 detail the quantities of waste generated and the total area
of disturbance on the mining lease.
4.2 AMD Potential
Static test work showed that both the oxide waste and sulfide waste have very low Acid Neutralising
Capacity (ANC). The ANC ranged between 0.0 and 18 kg H
2
S0
4
t-' with an average value
of 3.7 kg H
2
S0
4
r'.
AGC (1989) stated, "The static test work indicated that the oxide waste was naturally acidic
with low sulphur and low metal concentrations. Significant generation of acidic metal bearing
leachate was unlikely", and "The sulphide waste has a significant net acid producing potential
but although the more environmentally sensitive metals (e.g. arsenic, zinc, copper) are present,
the concentration of these metals are not as high as those for other projects reported in the
literature where AMD has been of concern."
45
TREDINNICK & CORNWELL
AMD Management at tho Pajingo Gold Mina.
I gl I To I IMouy w\ 0'""' [ .
,0 / I 0 8
__- '- 1D 0 .r--' N
/ I g
= ( . I .. --- ..
:I :
''""/ t 7- J-- Tt--f1-
--- I r. / \. J f ...
( ' __/ I /
I ; I /T1 / v
""'T/" "/'\'"' \ I j ' / /IHOWNG
I I \ ( . I SILTTRAPS 1 "
/ ,, .. )VIv, ,.,---
sEEPAGE RECLAIM ;: I ,r,t M I
..... ....._".. <.(
1
DAM oAM .. .1 oun
Tl . );;; Molly Dorl1ng L .
...--- . "'r' \"'GAZINE DUMP _ ''jni'A LODE
1
w ... .. <(,;, .. . '> l1'n. I : " DAM
--- '*',I .': . . ""'1!\1 :' . SAMS _______
1
'e.'\ TAILINGS ' '')\1"' '"
30000N , '-.. J ROCKORlL
- L .DAM . --i@j "" 00. IJCAMP
".., .. 2:?3 S\\ LOW E;l PlT ..::::::::::-
- ".,

j'-1/ ) \\;RADE
1
\ ...
-.:::::.-- _ -== -:::::7 , '\DUMP CAMP -....:;:'::-._
"'iN? ,.z,.fs ,.o;o /dfidNTRAcroRs'';.,oRKsHoP v .. / I --==:.
'\. ""'-"' I I '' :i ".,.
: IOFF.ICE,MlLL h//V / -. ) /'.:
.. ... .y/' / . i
2'>000N t -. =--)"''\\ I
- . - . \\ ? ROAD / ----...-
....... 1 .. CINDY WASTE
' \ \ ... .,,,, ., ....---
- "ht
11
' ,.
11
\ RADIO SHACK
,,,,, ' ";:: ---
1
"'" .,,\ , \ ''IJ.Ill' @J /Sil:'T DAM ,
', --- ,,,,, < 7'0- -- J
--- ""tf', ."' --. : DAM CINDY PIT
/'!''''''' -
---- I . ,. 11,__ _
2B000N I 1 i. ... / / ..,,\ -,r' __../' _J
/ -l-J r. rs.,; _,.---- .\ J
/-- .- .. \t- _/
27000N t ::' . .,//,. . --......__ __/ OAM

/ OOONGARA HOLDING
-.
1
/ti. Mount Ross .
.....---- - ...-......... FOUR
--- - --- ' '- \ --------
=-=-4 J Jrl\" .. /
7726000mN A.M.G \ "\ . _/ 0"---- ::>00 1000
I YO U.A H \{ rt SCALE 0 T! Top$011 Dump
440000mE A M.G 11000E 4ZOOOE I 43000E 44000E
I I f" \ ..,...-- I I

I
Fig. 1. Layout of Pajlngo Gold Mine showing areas disturbed by mining operations.
TREDINNICK & CORNWELL 46
Table 2. Waste generation.
Source
Scott Lode oxidised waste
Scott Lode sulfide waste
Cindy oxidised waste
Cindy sulfide waste
Tailings (oxidised and sulfide)
A Potentially AMD generating.
AMD Management at the Pajtngo Gold Mine
Waste generated
9 million tonnes
6 million tonnesA
1 million tonnes
20,000 tonnes
1.5 million tonnesA
Table 3. Mining lease disturbance (see Fig. 1).
Source
Scott Lode pit
Cindy pit
Scot Lode dump
Cindy dump
Tailings dam
Water catchment dams
Stockpiles
Treatment plant and offices
Haul roads
Exploration areas
Contractors yards
Topsoil stockpiles
Total
A Require AMD management or investigation.
Area disturbed (ha)
10A
3A
28A
BA
41A
11
4A
5
9
8
2
130
To further evaluate the potential leachate quality produced by sulfide waste, a series of column
leach tests were carried out Column leach tests were carried out on samples of sulfide waste
for a period of 31 months. The column test confirmed the acid-generating potential of the sulfide
waste.
The Tailings Dam facility is divided into three cells. Cells 2 and 3, and the base of Cell 1 have
had Scott Lode oxide tailings deposited into them. The middle layers of Cell1 will contain sulfide
Scott Lode and Cindy tails while the top layers will consist of Cindy oxide tails.
47
TREDINNICK & CORNWELL
li
I i
,,
AMD Management at the PaJingo Gold Mine.
5.0 PERSONNEL
Day to day responsibilities for environmental management fall under the mining department
(survey, geology, engineering). The mining department has consisted of between five and nine
personnel. Environmental responsibilities are the domain of the Senior Mine Geologist. Currently
this would account for 80% of his time. The limited available personnel has meant that operational
co-ordination and planning, and management of consultant's activities has been the key
environmental role for personnel on-site.
It was recognised early on that the operation was not large enough to support a dedicated
environmental department with sufficient expertise to assess the potential AMD issues and develop
suitable management plans. The use of Consultants was seen as a logical means of accessing
a wide range of expertise that the operation could not provide in house. Through exposure
to consultants, site expertise has been developed. This was further advanced through corporate
support of Post Graduate Environmental Studies.
Technical guidance is also provided by the Director, Environmental Services based in Denver.
As many of the policy changes in Australia over the last five years are copying the US system,
the 'inside brief' is a useful asset.
6.0 SCOTT LODE DUMP DESIGN
Early testwork in 1988 and 1989 indicated that the sulfide waste rock would need to be specifically
treated to ensure no ongoing legacy. Uterature searches, and pro-active mining practices indicated
that a careful encapsulation plan would be a practical cost-effective solution, if the correct material
was available.
Using exploration drill hole data, the oxide sulfide interface was generated on a digital terrain
model and overlain on the final pit design to evaluate the quantity and extraction schedule.
The principal features of the dump are:
(1) Basal oxide waste rock layer;
(2) 20 m thick (minimum) encapsulating oxide walls;
(3) Sulfide waste placed within oxide walls;
(4) Minimum of 3 x 0.5 m thick compacted cap layers;
(5) 0.5 m loose rock mulch surface layer;
(6) Contoured top surface to a central discharge point; and
(7) Monitoring of dump physics, surface runoff (quantity and quality), vegetation, erosion.
7.0 OPERATIONAL AND PLANNING CONSIDERATIONS
The principal operational and planning considerations involved in implementing the waste rock
encapsulation program were: contract negotiation, mine planning/scheduling, liaison with earth
moving contractor, effective control of dump construction, differentiation and segregation of waste
rock. The initial pit design consisted of an interim oxide pit from which to commence production
while final evaluation of the resource was carried out. The final dump design was not put in
place until after the open-cut mining contract was in place. To avoid additional haulage cost,
the limits of the dump were designed to be within 2 km of the pit centre. Post-contract negotiations
agreed that there would not be significantly more (or less) earth moving works required to construct
the encapsulating dump. No additional earth moving costs were incurred until the final capping
of the dump was undertaken.
TREDINNICK & CORNWELL
48
AMD Management at the Pajingo Gold Mine.
The Scott Lode pit at surface consists of a single pit which bifurcates into an East and West
pit approximately 60 m below the surface. A schedule by level of quantities of material required
to construct the waste dump was overlain on the mine extraction plan and model of material
types in the pit. Creative scheduling meant that the required quantities of oxide and sulfide
waste to construct the dump to the design parameters could be mined without the need to stockpile
and rehandle material.
By completing the East pit in advance of the West pit, it was possible to dispose of sulfide bearing
waste in the East pit. This had the advantages of enabling the dump to be capped one wet
season earlier and saved on waste rock haulage costs. The exposed sulfides in the pit walls
meant that disposal of sulfide bearing waste in the east pit would have no additional detrimental
impact on pit water quality on decommissioning.
The only material requiring rehandling was oxide waste lor the dump cap. This material was
stockpiled on sections of the dump completed earlier on. The maximum haul distance was kept
to approximately 350m. Towards the end of the mining of the Scott Lode pit, the mining fleet
was reduced from two to one excavator. The final benches of the pit were extracted by campaign
drilling and blasting followed by campaign excavation. By scheduling the capping operation
it was possible to utilise equipment that would otherwise have been idle and incurring standby
costs. No additional equipment was required to be mobilised.
The ability to differentiate oxide waste from sulfide waste visually meant that additional analysis
was not required. Waste segregation was incorporated as part of the normal grade control
procedure. No additional resources were required to maintain the waste segregation or survey
control on the dump, but a considerable amount of vigilance and control was required at times.
Particular effort was put into training contract earthmoving personnel to identify the rock type
at the digging stage, and placing it in the correct location on the dump. If in doubt, call it sulfides!!
Each 5 m bench was mapped and compared to the predictive model. As always, there were
variations, but in general, the overall quantity estimates were close to reality.
The Pajingo mine site is located in the Janet Range. The Janet Range is a series of low rocky
hills surrounded by flat open grazing land. The revegetation philosophy was based on not creating
a focal point for grazing in what would otherwise be unsuitable land for grazing. As such, all
disturbed areas are returned to woodland similar to the surrounding environment and not planted
with pasture. Pasture is only used to initially stabilise drainage lines. The erratic rainfall and
high annual variation compared to mean rainfall values makes the use of tree seedlings
unsuccessfuL All areas are direct seeded with a tree seed mix consisting of up to 20 species.
AMD management of tailings has involved the controlled deposition of sulfide-bearing tails into
the middle layers of Cell 1 only. This material will be covered by oxide tails from the upper
sections of the Cindy underground. Investigations are currently under way to refine the AMD
potential of the tails, and a final decommissioning strategy.
8.0 COSTS
The main cost centres incurred in the development and implementation of the AMD management
plan included Research/Consultants, Earth Works, Revegetation, and Monitoring Equipment.
Although additional construction supervision and survey control were required during the oonstruction
stage, these duties were conducted by the mine technical group without calling on outside
contractors.
49
TREDINNICK & CORNWELL
AMD Management at the Pajingo Gold Mine.
As previously mentioned, the most effective method for the Operation to access expert advice
was via consultants. Despite the initial apparent high cost of consultancy services and analytical
charges, they were viewed as necessary. The costs, when viewed in terms of the potential
long term liabilities, the chance to maximise the use of favourable materials, potential savings
in earth moving costs, and the ability to plan from a position of knowledge, were considered
necessary and cost effective.
Earth moving costs specific to the construction of the dump were limited to the capping. The
capping consisted of a minimum of three 0.5 m compacted layers and a loose rock mulch layer.
To achieve the desired gradients in the drainage lines, up to seven 0.5 m compacted layers
were required. The capping exercise cost approximately $50,000 ha-'. All other earth moving
costs were incorporated as operational costs. Operational costs would not have varied if the
dump had been constructed as an encapsulating dump or as a 'normal' waste dump.
Revegetation consists of establishing a cover of native trees via direct seeding. Pasture is used
to initially s t a b i ~ s e drainage lines. The cost of supplies, minor earthworks and manpower for
revegetation works is approximately $1500 ha-'.
Monitoring costs predominantly consist of analytical charges; the majority of these being for
surface water quality analysis. Monitoring of revegetation and other compliance conditions is
carried out by site personnel and the cost incorporated as part of normal operating costs.
By partially isolating the sulfide tails under oxide tails in Cell1 only, it is hoped to minimise the
thickness of additional cover required to decommission the tailings dam. Managed deposition
during the final months of operation will partially form the final required contours, minimising
the earth works required.
9.0 MONITORING
The ultimate objective of any mining company is to be able to relinquish a lease. As such it
was apparent to BMAI that accurate monitoring would need to be done to convince the Regulators
that the objectives of the Closure Plan had been adequately met. Considerable numbers of
water samples have been collected and analysed at the pre-mining stage, during mining operations,
and now during rehabilitation. Particular care has been taken to 'instrument' the water balance
on the waste dump, namely rainfall, evaporation, runoff and infiltration. A conveniently located
weather station, v-notch weir, lysimeters and temperature probes provide the base data for
the water balance to be calculated. Vegetative cover and erosion monitoring are done via visual
inspection.
10.0 MONITORING RESULTS
Monitoring of surface water quality downstream of the Scott Lode dump indicates that the sulfides
have been successfully isolated. Results are well within the compliance levels. Revegetation
on the dump is well advanced and after 2 years has already reach the target levels. This growth
provides great encouragement, as this has been achieved during periods of record drought with
no artificial irrigation at all.
11.0 CONCLUSIONS
Government, community and indeed mining companies themselves no longer find it acceptable
to behave as the alluvial miners of Victoria in the 1850's.
TREDINNICK & CORNWELL
50
AMD Management at the Papngo Gold MimJ.
Operations must now comply with stringent environmental guidelines (which no doubt will get
more stringent), and less than enthusiastic general populace.
To continue to exist as a viable operator, it is imperative to take into consideration the environmental
issues as early as possible. Careful and pro-active planning, stringent implementation and sensible
monitoring should result in an operation setting excellent environmental standards, which should
lead to compliance with lease conditions and a trouble free relinquishment.
We at Pajingo are proud of our progress to date. The final verdict will be known in some years
time.
12.0 ACKNOWLEDGMENTS
The authors wish to thank the fellow members of the staff at Pajingo who contributed to the
preparation and review of this paper and to Battle Mountain (Australia) Inc. for permission to
write and present it.
In addition, the authors acknowledge A.G.C. Woodward Clyde's input into environmental
management at the Pajingo Mine and in particular Peter Ryan for his contribution to AMD
Management at Pajingo.
13.0 REFERENCES
A.G.C. (1989). Pajingo Gold Mine Preliminary Waste Characterisation Study (unpublished).
Blainey, G. (1991). Gold- A Lesson or Two? Luncheon Address at the Australian Gold
Conference, Burwood Convention Centre, Perth, 16th March 1989. In 'Biainey Eye on
Australia.' (G. Blainey.) (Schwartz Books.)
BMGC. (1993). Statement of Environmental Policy and Principals, BMGC 1993 Annual Report.
51
TREDINNICK & CORNWELL
52
Second Australian Ac1d Mine Dra1nage Workshop (Eds N J Grundon and L.C. Bell). pp. 53-65.
PRACTICAL ENGINEERING OPTIONS
TO MINIMISE AMD POTENTIAL
A. Watson
AGC Woodward-Clyde Pty Ltd, 49 Park Road, Milton, Old 4064.
ABSTRACT
The potentia/liabilities related to the occurrence of Acid Mine Drainage (AMD) within metalliferous
mines throughout eastern Australia are substantial. The management of the AMD problem is
an integrated process involving the key elements of identification and characterisation, quantification,
scheduling, and finally engineering for control. This paper identifies several engineering options
to minimise the occurrence, or control the effects of AMD within mine waste dumps or tailings
impoundments. Integrated strategies related to design and construction of new waste dumps
and tailings storages to mitigate AMD potential are investigated. Options for closure, rehabilitation
or remediation of existing dumps and impoundments to reduce AMD effects are also explored.
Finally, a summary of key engineering design issues related to each of the options is presented.
1.0 INTRODUCTION
The recognition and control of acid mine drainage (AMD) are key issues that currently confront
the Australian mining industry. The AMD issue is one that has received increased exposure
over past years, and therefore the industry is becoming more aware of the potential scale and
long-term environmental significance of the problem.
Against a background of increasing regulatory control (e.g. recently introduced Environmental
Protection Legislation in Queensland), the potential financial liabilities of AMD occurrences need
to be more seriously considered. These liabilities could be associated with up-front 'securities',
increased control for waste handling and placement, or site rehabilitation and clean-up. The
significance of these liabilities, given the public attention on environmental, land use, and water
quality issues, and now with substantial support from government legislation, can not be
understated.
The control of AMD is clearly at the hands of the industry; initially in the diagnosis of the problem,
but then in the implementation of appropriate control measures. To this extent, this paper provides
a broad overview of some practical engineering options to either minimise the potential for AMD,
or to control AMD effects.
The paper has been structured to investigate initially the engineering process related to
implementation of AMD control measures. Several engineering options for AMD control are
then described. Finally, a summary of key engineering design issues related to each of these
options is presented.
The problem of AMD is relevant to mine waste dumps, mine tailings impoundments, heap leach
pads, mine adits, open-cuts and other areas. This paper focuses on the control of AMD around
mine waste rock areas and tailings impoundments as potentially the most serious AMD-related
problem currently facing the industry.
53
WATSON
Options to Minimise AMO Potential.
2.0 THE ENGINEERING PROCESS
The fundamental principle to minimise AMD potential is the isolation of the source material from
either of the key ingredients of AMD- air or water. Current technology and practice (and associated
economics) indicate that the concepts of encapsulation or entombment (in accordance with general
landfilling principles) comply with these objectives.
The generic engineering process related to design and implementation of AMD control measures
includes such steps as materials characterisation and quantification, evaluation of
performance/design criteria, definition of the solution and performance auditing. This process,
shown in logic form on Fig. 1, can be applied equally to a proactive or reactive approach to
design.
The proactive approach would involve the fomnulation and implementation of controls in anticipation
of an AMD problem. The design of such preventative measures would typically be applied at
the onset of development of a new waste impoundment or containment area and would continue
throughout the life of the faculty.
-
-
-
-
-
-
-
-
-
-
r
Materials characterisation/guantification
Geochemical
Physical
I
Siting
Constraints
Physical site conditions
I
Design QhilosoQh'!'/Qerformance criteria
I
Potential interactions
Performance standards/criteria
I
I
Solutions design
Engineering analysis
Predicted performance against criteria
'
I
Economic evaluation
Feasibility assessment
Techniques for cost control
I
lmolemeo"lloo, mooltoclcg ~ d m""''"'"'' I
Cradle-to-grave philosophy
Actual against predicted performance
Ongoing performance
Fig. 1. The engineering process
WATSON
54
Options to Minimise AMO PotentiaL
Conversely, the reactive process would be initiated in response to the occurrence of AMD and
would be implemented as a form of remediation. The engineering process for appropriate remedial
measures related to AMD control would typically apply to the closure, rehabilitation, or clean
up of existing dumps and impoundments to reduce or control AMD effects.
The following provides a brief description of each step included within the engineering design
process.
2.1 Materials Characterisation and Quantification
Geochemical and physical characterisation of materials to be handled within the impoundment
or containment area would be as follows:
(1) Geochemical: Delineation between 'problematic' and 'benign' materials with respect
to AMD potential. This may involve identification of the threshold conditions
(e.g. %S) which may generate AMD.
(2) Physical: Definition of the relative quantities of material types and the schedule
under which these materials may be encountered or exposed.
This step would dictate the need to consider AMD control as a key criteria to design of the
impoundment or to confirm the occurrence of AMD.
2.2 Siting
The proposed siting (or current siting) of a waste rock dump or tailings impoundment is an integral
part to the engineering process in relation to AMD control. The sensitivity of background site
conditions to impact, or the beneficial uses of these conditions, are relevant in the selection
of appropriate design standards.
2.3 Design Philosophy and Performance Criteria
The design philosophy for a containment or impoundment area can be established tiy identifying
the potential interactions between the site and the facility, and assessing the tolerance to which
the site conditions (e.g. water quality) could be impacted without affecting some assumed beneficial
use, or without exceeding some adopted environmental standards.
The design philosophy would guide the selection of the performance standards or criteria to
which the facility or works, as appropriate, would be engineered.
2.4 Solutions Design
Based on the adopted performance criteria, the design of the containment or impoundment area
could be carried out by applying appropriate engineering techniques (e.g. computer modelling)
to predict the perfonnance of the selected solution against the adopted standards. Key engineering
considerations would include seepage generation quantities, water quality and stability.
2.5 Economic Evaluation
Following completion of design of the facility or works, an economic evaluation of the proposed
solution would be appropriate. This costing would link with either the project feasibility assessment,
or the consideration of end-of-project profitability.
55
WATSON
Options to Minimise AMD Potential.
This costing would be influenced by the strategy of implementation of the works. Techniques
to minimise overall cost would be to integrate the construction work with the existing mining
operation (e.g. use of run-of-mine waste as cover, and use of mine plant in construction).
Furthermore, as a means of reducing large capital outlay at the completion of the project, staged
implementation of AMD control measures should be considered.
2.6 Implementation, Monitoring and Maintenance
Full implementation of the proposed scheme would be based on a cradle-to-grave philosophy
in order to achieve the adopted performance criteria. The implementation would culminate in
the final decommissioning and closure of the containment/impoundment area, whether based
on the proactive or reactive process.
Monitoring and maintenance of the works would form an integral part of both the implementation
and post-closure care plan. Monitoring would be used to gauge actual performance against
the predicted performance and adopted performance criteria. Maintenance would be necessary
to ensure on-going performance through the phases of implementation and post-closure care.
3.0 ENGINEERING OPTIONS
A broad review of the engineering options to minimise AMD potential or to control the effects
of AMD is provided in the following discussion. Options for both waste rock dumps and tailings
impoundments are described. The options investigated are methods that have been tria/led
and found to be effective, and have proven to be, under certain circumstances, economically
feasible to implement. Table 1 summaries the options investigated.
Table 1. Summary of engineering options.
Design approach
Proactive:
Preventative
measures
Reactive:
Remedial measures
WATSON
Waste dumps
1. Encapsulation
2. In-pit disposal
3.
4.
5.
1.
2.
3.
Mixing/Co-disposal
Micro-encapsulation
Uncontrolled
placement with
downstream
recovery and
treatment
Covers
Downstream
recovery and
treatment
Removal
56
Tailings impoundments
1. Containment and
cover
2.
In-pit disposal
3.
Co-disposal
1.
Containment and
cover
2.
In-pit disposal
3.
Co-disposal
I
I
Options to Minimise AMD Potential.
3.1 Waste Dumps
3.1.1 Encapsulation
The encapsulation option would be applicable for an above-ground dump area. Conceptually,
benign (i.e. non-acid producing) materials, such as oxide mine waste, would be used to encapsulate
the AMD portion of the waste stream, thereby isolating the problematic waste from air and water.
Typically, a deep basal layer of benign materials would be formed (or otherwise reliance on
the containment propertes of the in situ foundation sequences be made) with perimeter containment
wings formed using similar materials. The cell formed would be utilised for placement and
containment of problematic waste. This surface of this cell could be covered using benign materials,
either in staged manner, or on closure of the dump. The design issues related to the concept
include:
( 1) Achieving low permeabilities within the encapsulation zones, particularly the surface,
which is subject to direct rainfall;
(2) Providing adequate surface water management in and around the encapsulation
cell to avoid periods of inundation of AMD-producing waste during the operation;
and
(3) Providing a competent foundation/basal layer to reduce seepage throughflows
and protect downstream groundwater or surface water resources. The use of under-
drainage beneath the impoundment, particularly on steep/valley slopes, could be
considered.
The encapsulation option is shown conceptually in Fig. 2.
3.1.2 In-pit disposal
The use of in-pit disposal would only be applicable provided that a void (e.g. completed open
cut) of sufficient volume was available. In many cases, this is not the case. However 'creative'
pit management within an existing open cut operation can enable allocation of air space for
placement of waste rock prior to completion of mining.
Similar design concepts to the encapsulation option would apply to in-pit disposal, including
provision of basal, perimeter and surface encapsulation layers. The requirement for basal and
pit wall encapsulation would depend on the permeability of these faces, and also on the
constructability of these layers. The thickness of the encapsulation layers would be determined
by the waste volume balance (problematic to benign waste from the mining area) and available
supplementary borrow sources, as required. The design issues related to this option include:
(1)
(2)
(3)
Achieving low permeability encapsulation zones on the base, walls and surface
of the AMD waste cell;
Providing adequate surface water management in and around the encapsulation
cell; and
Assessing the potential connection between the pit and the local groundwater system
and to ensure limited 'cross flow' from the encapsulation zone to the aquifer.
The in-pit disposal option is shown conceptually in Fig. 2.
57 WATSON
Options to Minimise AMD PotentiaL
! 1 !
/
CO-DISPOSAL
OXIDE I BENIGN
WASTE
NOTE: ARROWS INDICATE
POTENTIAL WATER
MOVEMENT
MIXTURE
Fig. 2. Diagrammatic representations of engineering options of waste rock dumps using
the encapsulation, In-pit disposal, co-disposal, and mixture methods.
3.1.3 Mixing and co-disposal
The principles of the mixing/co-disposal option are the blending or selective placement of
problematic waste with a benign or non-acid producing material. The AMP potential of these
wastes are reduced through a combination of reducing the net acid producing potential of the
waste, in addition to partial encapsulation.
This option is shown conceptually in Fig. 2, in two forms. The first form comprises the development
WATSON
58
Options to Minimise AMD Potential.
of small cells for placement of problematic waste. These cells would be formed from benign
waste or other suitable borrow materials. The second form would be the mixture of both the
benign and problematic waste streams. This method would rely more on the neutralisation capacity
of the benign materials than its encapsulation properties.
3.1.4 Micro-encapsulation
A methodology for coating of certain mine wastes has been developed to prevent pyrite oxidation.
The mechanism of this coating technology involves the leaching of mining waste with a phosphate
solution containing hydrogen peroxide. When this solution reaches pyrite surfaces, the hydrogen
peroxide oxidises the surface portion of the pyrite and releases iron oxides so that phosphate
precipitation forms a passive surface coating. It is understood that this process has not been
trialled on a large scale.
3.1.5 Uncontrolled placement with downstream recovery and treatment
A possible option for control of marginally acid producing waste is the uncontrolled or undefined
placement of the problematic materials, then use of downstream recovery and/or treatment of
any acidic drainage that may subsequently be generated. Possible recovery systems would
comprise recovery ponds, subsurface seepage drains, and groundwater recovery bores. Treatment
system options could include reagent treatment (e.g. lime dosing), wetland filter treatment, and
dilution in an adjacent water course. A key engineering issue related to this option is provision
of an appropriate surface waste management system leading to effective collection and/or
containment of drainage to facilitate treatment. Careful siting of the waste dump is therefore
critical under this option, to avoid the misdirection of acidic drainage. A typical representation
of this scheme is shown on Fig. 3.
3.1 .6 Covers
Applying the principles of encapsulation, a low permeability cover can be constructed over an
existing waste rock dump area to reduce the infiltration of surface water and to minimise the
infusion of air. The use of locally available borrow or suitable benign mine wastes are favoured
in forming these covers, largely on the basis of economics. Additionally, the cover thickness
would be minimised to reduce the overall earthworks requirements. The design process related
to this option would include detailed hydrological and seepage/mass balance modelling based
on the physical properties of the cover materials to attempt to optimise the required cover thickness.
Key design issues related to this option are:
(1) Assessment of requirements to include capillary breaking or barrier layers within
the cover to minimise salt rise or root penetration through the thin encapsulation
zone. The need for these layers would depend on the hydraulic behaviour of the
waste rock mass;
(2) Consideration of constructability of covers, based on design thickness and
slope/grade of waste dump surfaces; and
(3) Consideration of durability/serviceability with respect to long-term periormance
of covers. This would be related to desiccation/shrinkage cracking, weathering
and erosion potential. The use of surface protection layers would need to be
considered in this event.
This option is shown conceptually in Fig. 3.
59
WATSON
Options to Minimise AMD Potential. I Options to Minimise AMD Potential.
' ~
OXIDE I BENIGN
WASTE -...
~
y
COVERS
WETLAND
GROUNDWATER II ~
RECOVERY BORE c ~
NOTE: ARROWS INDICATE
POTENTIAL WATER
MOVEMENT
ELEVATION
DOWNSTREAM RECOVERY AND TREATMENT
Flgo 3, Diagrammatic representations of engineering options of waste rock dumps using
covers, and downstream recovery and treatment niethodso
3.1.7 Removal
The removal of waste rock materials is an option that would not be considered routinely. However,
some advantage may be assessed in selective mining of severe waste from an existing dump
for isolation either within an open cut, void, or an alternative, suitability prepared waste dump.
WATSON
60
3.2 Tailings Dams
3.2.1 Containment and cover
Typical tailings storage development within northern Australia comprises above-ground containments
designed for sub-aerial tailings deposition methods. The resulting structure would comprise
a perimeter (on flat sites) or containment (on valley- fill sites) embankment, with an exposed,
dried tailings surface.
The features of such a storage, included for AMP control, would be a low permeability core
or blanketing within the embankments, partial or full storage floor lining (typically earthfill
construction) and a low permeability cover formed over the final tailings surface. Typically, a
'starter' embankment and storage lining would be constructed initially, followed by staged lifting
of the containment to reach the design storage capacity. The surface covering would likely
be constructed following closure of the facility.
The key issues related to the design of this option is the detailing of the storage floor lining and
the surface cover. The use of locally sourced, clay-dominant soils is common in this application.
Alternatively, for the storage lining, synthetic membrane (e.g. HDPE) has been used in some
cases, principally for containment of particularly problematic waste, or lor protection of particularly
sensitive site conditions. (The cost implications and uncertain design life of this alternative is
significant).
For the surface cover, a number of additional design considerations apply and a number of cover
alternatives exist. The additional design considerations relate to the reduction of capillary rise
which would bring dissolved salts through any earthfill cover to the layer surface. This rise may
hinder vegetation growth. Similarly, root penetration through the surface cover would promote
downward migration of moisture which would recharge the AMD mechanism. A capillary breaking
layer or similar barrier layer within the cover may therefore be considered.
The alternatives for the cover. which could be applied with or without the capillary breaking barrier
layer as appropriate, would be synthetic membranes (e.g. HDPE), organic blankets, cementitious
covers, surface chemical methods (e.g. hardpan development), or wet covers.
Other design issues related to the development of this option include:
( 1) Surface water management, particularly diversion of storm flows around valley-fill
sites;
(2) Management of the underlying groundwater related to the reduction of seepage
losses from the storage; and
(3) Long term embankment stability. related to both slope failure protection and control
of erosion and scouring.
This option is shown conceptually in Fig. 4 for both a flat and valley-fill site.
The downfall of the described method of tailings storage development in relation to AMP control
is that typically, the containment layer cannot be stage constructed. This has implications in
long term AMP potential and in high rehabilitation costs to the end of the project.
61
WATSON
Options to Minimise AMD Potential.
OXIDE I BENIGN
WASTE
~
/
~
OXIDE I BENIGN
WASTE
'
DIVERSION
BUND
NOTE: ARROWS INDICATE
POTENTIAL WATER
MOVEMENT
Fig. 4. Diagrammatic representations of engineering options of tailings dams using the
containment and cover method.
3.2.2 In-pit disposal
Similar to the in-pit disposal option for waste rock placement, such a scheme for tailings disposal
would only be appropriate in the event that a decommissioned or unused void of adequate storage
volume was available. Due to the difficulties of an encapsulation zone within the open pit, the
in situ low permeability of the pit walls and floor would typically be relied upon to enable adequate
containment. A surtace cover would be constructed on closure of the system. The design issues
related to this cover and appropriate cover alternatives would be as discussed in the Containment
and Cover option (see 3.2.1 ). The key design issues related to the in-pit disposal option are:
( 1) Assessing the possible connection between the pit and the loca: groundwater system;
and
WATSON
62
IN-PIT DISPOSAL
Opt1ons to M1n1mise AMO Potential
NOTE: ARROWS INDICATE
POTENTIAL WATER
MOVEMENT
t . . " ' ' < : : { ~ ' ' " '
P//htt: ( f ( &!&V?7b.-
Fig. 5. Diagrammatic representations of engineering options of tailings dams using the
In-pit disposal and co-disposal methods.
(2) Providing adequate surtace water management within the void to prevent long
periods of inundation of the tailings.
This option is shown diagrammatically in Fig. 5.
3.2.3 Co-disposal
The concept of co-disposal varies between the various mining industries. Related to coal mining,
a co-disposal system typically describes the mixture of the fine and coarse reject fractions.
More detail on this method is provided in Williams (1992). Related to metalliferous mining, a
not too widely used technique is the co-disposal of mine waste (rock) with process tailings.
This method would involve creation of a series of tailings cells using stable and suitably non-acid
63
WATSON
I
'
~ ~ ~ ~
I'
I
II'
'li
,1::
I'. I
li
1
1
1
iii
li;
~
I H\!
Options to Minimise AMD Potential
producing waste rock. Conceptually, each cell would comprise a relatively small volume to enable
rapid filling, followed by a period of drying, then covering. Sequential cells would be developed
to meet the process requirements. The specific requirements of the system would be the use
of low permeability materials to form each cell, and reduced cell volumes to control water
managements aspects and to minimise ongoing seepage potential. This option is shown
conceptually in Fig. 5.
4.0 SUMMARY OF KEY ENGINEERING ISSUES
On the basis of the engineering options described in the previous sections, a summary of key
engineering issues has been compiled. This summary provides a general list of issues, not
all issues being relevant to the one option:
(1) Providing adequate characterisation/quantification of the waste stream.
(2) Achieving desired permeability within waste encapsulation zone.
(3) Providing adequate water management system. This may include:
(a) Surface water- diversion of surface water around the site, or containment
of surface water, as appropriate;
(b) Groundwater - mitigation of seepage into adjacent groundwater systems,
or alternatively recovery and treatment of seepage; and
(c) Unsaturated zone use of capillary breaking or barrier layers to limit salt
rise or downward root penetration.
(4) Ensuring long term stability of the waste containment areas.
(5) Undertaking appropriate monitoring and maintenance of the system to assess ongoing
performance and to ensure long term serviceability.
(6) Use of appropriate techniques and construction materials to provide an effective
and economical solution which is stable and non-polluting.
It is further emphasised that the direct cost and related financial liability of AMD control measures
is a significant engineering design issues for waste dumps and tailings dams. Intuitively, many
advantages are foreseen with taking a proactive approach to AMD control, with respect to staged
development of the works and the effective reduction/control of the AMD process.
5.0 REFERENCES'
Phinney, K.D. (1992). Planning for acid generation : An engineer's perspective. Second
International Conference on Environmental Issues of Waste in Energy and Mineral
Production, Calgary, Canada.
Vick, S.G. (1983). 'Planning, Design and Analysis of Tailings Dams.' (John Wiley and Sons:
New York.)
1
Editors' Note: These references have not been cited in the text. They are included at the
author's request to provide background information relevant to the subject of the paper.
WATSON
64
Options to Minimise AMD PotentiaL
Ryan, P.A., and Joyce C.J. (1991). Prevention and remediation of acid drainage from mine
waste rock in Australia. Second International Conference on the Abatement of Acidic
Drainage, Montreal, Canada.
Ahmed, S.M. (1994). Surface chemical methods of forming hardpan in pyrrhotite tailings and
prevention of the acid mine drainage. Third International Conference on the Abatement
of Acidic Drainage, Pittsburgh, PA.
Meek, F.A. (1994). Evaluation of acid prevention techniques used in surface mining. Third
International Conference on the Abatement of Acidic Drainage, Pittsburgh, PA.
Williams, D.J. (1992). Emerging techniques for disposal of coal wastes. Second International
Conference on Environmental issues and Management of Waste in Energy and Mineral
Production, Calgary, Canada.
65
WATSON
66
CONTROL OF ACID MINE DRAINAGE THROUGH WATER
MANAGEMENT AT MT. LEYSHON GOLD MINE
M.S. Orr
Normandy Poseidon Group, P.O. Box 6080, Townsville Mail Centre, Old 4810.
ABSTRACT
The issue of long-term acid mine drainage at Mt. Leyshon Gold Mine is being addressed via
a waste dump rehabilitation strategy comprising selective waste handling and dump sealing.
However, during the mine life, there will always be unsealed areas of active waste dump, and
also ore stockpiles. To minimise the impact of these areas on the quality of local surface water,
the mine has implemented a site runoff control strategy. This comprises a system of diversion
weirs, dams, and sumps that achieves containment and treatment of all flows, but allows a
proportion of high flows to safely leave the site. The system includes continuous monitoring
of the duration, volume, and quality of discharge.
1.0 INTRODUCTION
Mt. Leyshon Gold Mine (MLGM) is located 24 km south of Charters Towers, in northern Queensland
(Fig. 1 ). It is a large open-pit mine, producing gold, silver and copper. The operation is owned
by Mt. Leyshon Gold Mines Limited, which is a member of the Normandy Poseidon Group, a
predominantly Australian-owned mining finance company.
Mt. Leyshon Gold Mine began operations in 1987. Operations were initially restricted to the
heap-leaching of oxide ore. Ongoing testwork revealed a much larger resource of unoxidised
ore at depth, which lead to the construction and operation of a conventional carbon-in-pulp (CIP)
plant. Throughput is currently 5.5 Mt year'. Heap-leaching of oxide ore has now ceased due
to depletion of the resource, but a heap-leach process is now being used to treat copper-enriched
supergene ore.
2.0 SITE FACTORS
2.1 Climate
Mean annual rainfall at Charters Towers is 660 mm, with most rain occurring from November
to March. An important feature is the great variability in rainfall, from year to year and from
location to location. Much of the rainfall is typically received in high intensity storms associated
with high rates of runoff. Droughts are common. Average pan evaporation is 1,965 mm (Fig.
2). Average maximum and minimum temperatures for Charters Towers are 30 oc and 17 oc
respectively.
2.2 Geology
The Mt. Leyshon orebody is hosted in an intrusive breccia and igneous complex approximately
1.6 km in diameter. Polymetallic mineralisation (Au-Fe-Cu-Zn-Pb-Ag-Mo-Bi) is hosted in
breccias and associated intrusives. The main host to mineralisation is the MI. Leyshon breccia,
a roughly pipe-like body enveloped within the larger (predominantly barren) Main Pipe Breccia.
A second host, the Mine Porphyry, accounts for approximately 15% of the mineralisation.
67
ORR
Control of AMD Through Water Management.
ORR
c:-J
\j
o 10 20 XI raan
L..._l 1......,1 '---..1
Fig_ 1 _ Location of Mt Leyshon Gold Mine.
E
E

200
150
100
5
]111111 b L L lll L B
Jan March May July Sept Nov
Feb April June August Oct Dec

pan evaporation Mean



Fig. 2. Average rainfall and evaporation for Charters Towers.
68
Control of AMD Through Water Management.
2.3 Solis
The soils on the tenement areas are typically stony and shallow(< 0.5 m). Apart from limited
areas of sandy alluvium in waterways, red and yellow duplex soils are dominant. Weathered
dolerite dyke outcrops yield small areas of self-mulching yellow and brown clays. Quartz porphyry
dykes yield local sandy deposits. The soils are non-saline and nutritionally limiting.
2.4 Vegetation
The natural vegetation is mainly dry sclerophyll woodland. The dominant species are Eucalyptus
crebra (narrow-leaved ironbark), E. melanophloia (silver-leaved ironbark), E. papuana (ghost
gum), Acacia argyrodendron (blackwood), A. bidwillii ( corkwood), A. holosericea and A. salicina_
Many years of cattle grazing have resulted in the removal of most of the understorey vegetation.
Ground cover is dominated by Heteropogon contortus (black spear grass) and Cymbopogon
refractus (barbed wire grass). The weeds Ziziphus mauritiana {chinee apple) and Cryptostegia
grandiflora (rubber vine) are common on disturbed areas in the district
2,5 Topography
Landforms at the site generally comprise alluvial flats and undulating hills and ridges. Examples
of the later include Mt. Hope (444 m) and Mt. Leyshon itself, which before being mined was
the highest peak in the area (522 m).
2.6 Water Resources
The lease areas are drained by many gullies that flow northwards towards Clarke Creek or
southwards to Puddler creek. Both Puddler Creek and Clarke Creek join the Broughton River,
which flows into the Burdekin River some 30 km from the mine site. The Burdekin River runs
continuously for most years, but the lesser waterways mentioned above flow only during or
immediately following rainy periods. Local groundwater reserves are sparse and are generally
only found in fractured zones and structural features associated with volcanic rock_ Local windmills
are typically used for stock watering, and yield low volumes of shallow groundwater from
watercourse underflow and near-surface fractures.
3.0
RUNOFF CONTROL MANAGEMENT
3.1
The Need for Runoff Control Management
As commonly found at many metalliferous mines, the primary (unoxidised) waste rock and ore
at MLGM contains varying levels of sulfides, mainly in the form of pyrite. When exposed to
oxygen and water, these minerals will oxidise to produce sulfuric acid. Sufficient water passing
through this rock will result in metal-rich seepage exiting the dump or stockpile, commonly known
as acid mine drainage. The rehabilitation strategy for the waste dumps at MLGM is therefore
designed to minimise the potential for acid mine drainage. This strategy is discussed in a later
paper in these proceedings (Orr 1995).
However, during mine operation, there will always be active, unsealed areas of waste dump
and active ore stockpiles. To control the impact of these potential sources of acid mine drainage
on surrounding waterways prior to rehabilitation or processing, the mine has implemented a
runoff control strategy.
69 ORR
Control of AMD Through Water Management.
3.2 The Runoff Control System
Total containment of all flows from catchments containing open waste dumps and ore stockpiles
at MLGM would require the provision of very large storage dams and incur the risk of a net
positive water balance for the site. It was therefore decided to implement a system that allows
partial release when conditions are suitable. This system contains two main components, namely,
runoff control dams and runoff control sumps.
3.2.1 Runoff control dams
3.2.1.1 General
There are four runoff control dams on-site, located in gullies draining waste dumps and stockpiles
(Fig. 3). Their operation is summarised as follows:
1. At low to medium flow rates, acid mine drainage heading off-site is entirely contained
by the runoff control dam. All water contained by the dam is restored to neutral pH
through slaked-lime addition. This process precipitates copper, allowing the water used
as process water with no detrimental effect on cyanide consumption in the CIP process.
The treated water is transferred to the plant or the site water storage dam as soon as
practicable.
2. At high flow rates, such as seen during severe storms, a proportion of the runoff is
automatically bypassed around the runoff control dam, to leave the lease. The diversion
weirs that achieve this shedding of water are sized so that clean runoff from adjacent
catchments is sufficient in volume to dilute contaminants to below potentially harmful
levels.
D ~ u t n g
clean runoff
Hghllow
da:harge
Acid runoff
Fig. 3. Configuration of a runoff control dam at Mt
Leyshon Gold Mine.
ORR
70
3.2.1.2 Diversion weirs
This automatic bypass during
extreme rainfall events is achieved
by a double weir arrangement located
at the upstream end of the dam (Fig.
4). Flow through both weirs (one
leading to the dam, one leading off-
site) is a function of water height and
weir dimensions. The weir sill height
determines when flow starts, while
the width determines the quantity of
flow, as described by the relationship
shown. The performance of this weir
system can be readily modified on
the basis of monitoring, for example,
using sleepers to raise the sill height.
3.2.1.3 Water treatment
The runoff reaching the runoff control
dams is treated with slaked lime as
required to restore pH to around 6.5.
This is sufficient to remove most of
Height data logger and
rising stage sample1 Slaked lime
+
To lease boundary
To runoff control dam
3/2
Cippoletti weir: Flow ::: 0.567 x L x H
Fig. 4. Weir configuration at entry to the
runoff control dams.
Control of AM 0 Through Water Management
the copper from solution. The lime slaking
plants were designed and built on site, and
can each apply up to two tonnes of lime
per hour through a screw feeder and
mixing tank (Plate 1 ). The slaking plants
are situated to discharge directly into the
runoff stream as it enters the dam,
resulting in good mixing. An important
feature is that dosing takes place in the
dam itself, eliminating scaling problems
in vessels or pipes. Operation of the
plants can be completely automatic, with
dosing rate proportional to rate of runoff
flow (via an ultrasonic level probe).
Plate 1. View of a lime slaking plant and double weir system at one of the runoff control
dams.
3.2.2 Runoff control sumps
Runoff control sumps are used in place of runoff control dams in areas where the catchments
are changing from year to year, namely, in front of advancing waste dumps. The sumps are
relatively low cost installations. The use of skid-mounted pumps and surface-laid HOPE pipe
allows easy relocation of the sump before each wet season (Plate 2). The function of a sump
is to transfer runoff to a runoff control dam. Consequently, the sumps have no provision for
water treatment and no diversion weir system. Pump operation is automatic. The sump locations
are moved progressively ahead on the advancing dump. Each is sized to cater tor a 1-in-1 0
year storm event, based on the maximum areas of cleared ground and open dump to be serviced
by the sump.
71 ORR
Control of AMD Through Water Management
Plate 2. View of a runoff control sump.
3.2.3 Integration of the system
The runoff control dams and sumps operate together to form a site water management system
(Fig. 5). The operation of this system during storm periods requires detailed surveillance and
recording. At the beginning of each wet season, the site's raw water storage is allowed to drop
down to provide a spare 400 ML of storage, in the event that storage of treated runoff will be
required. Treated water from the Mt. Hope dam is returned via the southern tailings dam decant
system, to avoid expensive replication of pipework. However, it should be noted that the system
avoids the use of tailings dams for storage of surplus water (water transferred to the southern
tailings dam is pumped back to the plant as soon as practicable).
3.3 Design of the System
The runoff control system was designed by Mcintyre and Consultants (a Townsville-based
engineering company) with strong involvement by MLGM personnel. The design process can
be summarised as:
ORR
(1) Calculation of catchment areas (clean and contaminated) at each runoff dam and
at lease boundaries;
(2) For different storm events, calculation of the volume of bypass allowable at each
dam to meet water quality criteria at lease boundaries (taking into account times
of concentration);
(3) Calculation of required dam and pump capacities;
(4) Modelling of site water balances for the scheme (for wet, dry and average years);
and
72
Treatment plant
... ,.,._ Automatic bypass
-+- AMD
Control of AMO Through Water Management .
......
Fig. 5. The site water management system.
(5) Engineering of diversion weirs to achieve design performance.
The design process was highly iterative, with multiple simulations being conducted using a
spreadsheet-based site water balance model.
4.0 MONITORING
To allow validation and 'fine tuning' of the site's runoff control system, and to satisfy the reporting
requirements of licence conditions, the following monitoring is conducted at each of the runoff
control dams:
(1) Water quality at different levels of flow (using rising stage samplers); and
(2} Change in water level over time (using water level sensors and dataloggers). The
automatic recording of water level at the diversion weirs located at each dam allows
calculation of the duration and volume of discharge from the system.
In addition, rising stage samples are located at the exit points of creeks from lease boundaries.
Numerous additional water samples are collected manually during rainy periods.
5.0 COSTS
The cost of the original runoff control scheme (design and implementation) was approximately
$1 million. Annual extensions (relocation of sumps) cost around $40,000.
6.0 REFERENCES
Orr, M.S. (1995). Development of a waste dump management strategy at Mt. Leyshon Gold
Mine. In 'Proceedings of the Second Australian Acid Mine Drainage Workshop', Charters
Towers, Queensland. 28-31 March 1995. (Eds. N.J. Grundon and L.C. Bell). pp. 91-100.
(Australian Centre for Minesite Rehabilitation Research: Brisbane, Australia).
73 ORR
74
Second Australian Ac1d M1ne Ora1nage Workshop {Eds. N.J. Grundon and L C Bell), pp 7 ~ - 8 2
MANAGEMENT OF AMD AT WOODLAWN MINES
S.E. Tarlinton
Woodlawn Mines, P.O. Box 141, Goulburn, NSW 2580.
ABSTRACT
Acid Mine Drainage at Woodlawn was not anticipated to be a problem for environmental
management. The EIS tor the project took sulfide oxidation in waste rock and tailings into account
but greatly underestimated its eventual severity. Within a year of commissioning, the site was
experiencing unexpected problems with water management and AMD control. An extensive
program of catchment minimisation and the promotion of evaporation through the 1980's failed
to reduce the volume of contaminated water stored. Cessation of open pit operations allowed
the rehabilitation of the waste rock dump. The dump was sealed by surface compaction, covered
with clean fill and vegetated. This rehabilitation, combined with further dam building put the
site into a negative water balance. The flow and concentration of leachate from the rehabilitated
dump have gradually decreased since completion of the seal. Rehabilitation effectiveness and
the evaporation of residual water will remain management issues following mine closure.
1.0 INTRODUCTION
Woodlawn Mines operates a base metals operation located in the Southern Tablelands of New
South Wales, midway between Canberra and Goulburn. The area has an annual rainfall of
around 700 mm and an annual evaporation rate of around 1400 mm. The elevation of the site
is from 750 to 850 m above sea level. The main land use in the district is the grazing of sheep
and cattle.
Drainage from the site is split between the catchments of the nearby Lake George to the west,
and the Wollondilly River in the east. Both of these catchments are sensitive to water pollution,
as Lake George has no outflow, and the eastern catchment feeds into the storage at Lake
Burragorang, the main water supply for the city of Sydney.
Ore production began in 1978, with open-cut mining at a rate of approximately one million tonnes
of ore per year. Underground production commenced in 1987, and production has exclusively
been from underground since 1989. The economic minerals produced are sphalerite (ZnS),
chalcopyrite, (CuFeS) and galena (PbS). Both the ore and associated host rocks contain significant
amounts of pyrite (FeS,). The tailings produced by the concentration process contain unrecovered
economic minerals and pyrite. Acid drainage from these materials has been the cause of most
problems in the environmental management of the Woodlawn site.
The water quality limits for runoff from the Woodlawn lease are those set down in Schedule 2
of the Clean Waters Act of New South Wales. This schedule in effect specifies the limits for
drinking water quality. In order to meet these limits, there must be no discharge of water from
the mine, tailings dams or plant area. The history of the control of Acid Mine Drainage (AMD)
at Woodlawn has largely centred on the storage and evaporation of acidic waters and the reduction
of sulfide-contaminated catchments (Hammond 1981; Sing line 1992; Skrinde 1979; Southern
1987).
75
TARLINTON
Management of AMD at Woodlawn Mines.
2.0 EARLY AMD CONSIDERATIONS
The original Environmental Impact Statement for the Woodlawn project was produced in 1973
(Jododex Australia Ply. Limited 1973). This document was amended in 1976 to take account
of the recently introduced Commonwealth and State environmental legislation (Jododex Australia
Ply. Limited 1976). Both of these documents consider the eventuality of some oxidation of sulfide
materials in waste rock and tailings. The concept of an effect on sulfides by Thiobacillus is
mentioned several times in the report. The severity of the AMD problem was, however, grossly
underestimated. The section of the EIS relating to the waste rock dump claimed that the water
percolating through the dump would never contain as much contamination as there was in the
existing groundwater around the orebody. The existing groundwater over the orebody contained
zinc concentrations of between 50 mg L' and 90 mg L'. Within the first year of operation,
the waste dump was yielding water with zinc concentrations of over 100 mg L'.
In the planning for the project, much faith was placed in the process of cation exchange. The
geochemical anomaly which led to the discovery of Woodlawn was created by the adsorption
of metal ions on clays. Accordingly, it was believed that metal ions liberated into the groundwater
around the tailings dam and waste rock dump would be removed from the leachate through
adsorption to clay particles in the soil. Although this process probably did occur to some extent,
its overall effect was negligible. The eventual metal load in leachate streams was far in excess
of the adsorption capacity of the highly dispersible local soils.
The plans for the construction of the waste rock dump included the classification of waste according
to sulfide content. The waste was then to be preferentially placed on the dump to minimise
the exposure of material with a high sulfide content to the atmosphere. The complex geology
of the orebody prevented this procedure from being managed effectively, and it appears to have
been abandoned early in the life of the project. At the completion of the dump, there was sulfide-
bearing material in abundance over most of the dump's surface.
Woodlawn's tailings storage was not anticipated to have been a source of contaminated water.
Management of the tailings discharge and alkaline addition in the flotation process was expected
to prevent an acid environment from developing. The first rehabilitation plan for Woodlawn tailings
involved the planting of local native plant seedlings directly into the dewatered tailings. The
need to store contaminated water from other parts of the site in the tailings dam, as well as
acid generation in exposed tailings, precluded the possibility of such revegetation.
3.0 WATER MANAGEMENT
Within the first year of operations at Woodlawn, water management had become a critical issue.
Initially, all of the circuit water for the mill was imported from an offsite borefield. It was quickly
recognised that this input represented unsustainable additional volume to be stored on site.
The available storage volume in the tailings dam was diminishing much faster than expected.
The first tailings dam was designed to operate for at least three years, but was 55% full after
eight months of operation. This was blamed on higher site water use than expected, and lower
evaporation.
The requirement by authorities that tailings be discharged below the water level in the dam
was blamed for the lower than expected settled density of the deposited slurry. This was also
adding to the rapid loss of storage volume. Tailings deposition was allowed to form a beach
from 1980, thus allowing the material to oxidise. This did not have an immediate effect on water
quality, as although Woodlawn tailings contain up to 25% sulfur, they may require up to six months
of exposure before the onset of acid conditions (Miller 1993).
TAR LINTON 76
Management of AMD at Woodlawn Mines.
Also in 1980, a water circuit to return supernatant water from the tailings dam to the concentrator
was established to minimise the need for further importation of water and a second, larger tailings
dam was commissioned. By that time, the quality of leachate from the waste rock dump was
deteriorating to the point at which specified water quality limits were not being met at the lease
boundary. This necessitated the raising of the spillway on the waste dump catchment dam and
the total containment of dump leachate and runoff. Water from the waste rock dump was pumped
into the tailings dams. In combination with the effects of tailings acidification, this input led to
a rapid deterioration in water quality. Returned water for the mill eventually required treatment
before it could be used in the metallurgical process.
3.1 Water Reduction Measures
Low rainfall in the early 1980's allowed the two tailings dams to store all the contaminated runoff
and leachate from the site. In order to maximise the loss of water, a system of spray lines was
constructed on the beached tailings. These were estimated to increase evaporation by up to
10% when operating at full capacity, but were plagued by problems of high maintenance
requirements. The waste rock dump was also utilised to provide additional evaporation of water
from the site. A series of ponds and sprays were installed on the upper levels of the dump,
but practical difficulties in preventing infiltration reduced their effectiveness. An 85 t haultruck
was converted to a water cart for dust suppression on the site roads, and this proved a bonus
for reducing contaminated water stocks.
These measures were inadequate for reducing the amount of contaminated water on the site
to a manageable level. A much greater evaporative surface area was required to bring the
site towards a net loss of water in a year of average rainfall. In 1987, the first evaporation
dam was constructed. Although this provided an additional 30 ha of evaporation area, it also
became a contaminated catchment of slightly larger size. The overall advantage of such a new
dam is thus reduced, and variations in the ratio of rainfall to evaporation can reduce its effectiveness
dramatically.
In the following few years, record rainfalls in the area brought about the need to raise the wall
of the second tailings dam, construct a third tailings dam and a second evaporation dam. These
developments and a fortuitous easing of rainfall allowed the company to narrowly avoid a major
spill from the tailings and evaporation dams.
A water treatment plant was trialled and commissioned in 1988-89, while the dam building program
was under way. It was intended that this would neutralise contaminated water and produce
water of a quality suitable for discharge from site. The effectiveness of this plant was overestimated,
and it was abandoned after only a few months of operation.
4.0 WASTE ROCK DUMP REHABILITATION
When mining from the open pit was terminated in 1989, the waste rock dump became available
for rehabilitation. At that time, the waste dump was the source of around one third of the
contaminated runoff on the site. The completion of this project would be the major factor in
bringing the site water balance under control.
Woodlawn's waste rock dump consists of between 60 million and 70 million tonnes of rock.
The anticipated sulfide content of the waste rock, as discussed in the EIS was an average of
1.3%. Bulk samples taken from .the dump after completion revealed a sulfide content of 6%
to 7%. The area of the waste rock dump when dumping finished was about 1 00 ha. It was
constructed in five levels to a total height of 80 m.
77
TARLINTON
Management of AMD at Woodlawn Minas"
4.1 Early Trials
The first rehabilitation trial had been conducted in the autumn of 1983. An area of 6 ha on the
lowest batter pushed down to a slope of 15to 20. Lime was applied to the rock surface at
rates of 4.0 t ha-' and 8.0 t ha-'. A nominal10 em layer of stockpiled topsoil was then placed,
and this was sown with suitable pasture species and fertilised. After a few months of encouraging
results, problems began to appear in the vegetation. At first small patches of grass began to
die, but these quickly developed into large sections of the batter. The pH of the soil had decreased
significantly, and investigations revealed that acid water from the dump was being drawn to
the surface by capillary action. The neutral ising capacity of the lime had been quickly overcome.
Later tests estimated that the lime addition required to neutralise acid generation to a depth
of 30 em would be in the order of 400 I ha-'. The failure of the first simple trial was not completely
unexpected, but it did confirm the need for a seal over the waste material to protect the vegetation
in the cover.
4.2 Design Development
The conventional rehabilitation technique of a layered cover including an impervious clay lining
over the waste was considered impractical for the Woodlawn dump. A conventio nal30 em layer
of clay would have required the provision of at least 300,000 m
3
of clay. Early in the life of the
mine, it was determined that such a source of clay would probably not be available on the site.
Rehabilitation of the clay extraction sites would also have significantly increased the rehabilitation
area on the site. The only available alternative was to use the waste rock itself as a compacted
sealing layer. The compaction of level haulroads by mine traffic was observed to allow water
to pool on the surface for weeks at a time. A series of experiments showed that compaction
in this uncontrolled fashion would effectively reduce infiltration to a minimum on most dump
materials. In 1983, studies by the Public Works Department and Soil Conservation Service
of NSW were commissioned to investigate the feasibility of applying this technique in the
rehabilitation of the dump.
The recommended level of compaction for a practical seal was 2.19 t m-
3
, which was achievable
on most of the material on the surface of the dump. A large vibrating roller was required for
the job, and this performed best when the dump surface had been smoothed with a grader to
improve contact with the roller. Allowing the material on the surface to weather for a period
of at least six weeks was also found to improve the efficiency of compaction.
Two rehabilitation trials were conducted in 1984 using the compaction method and a greater
depth of cover. These fared better than the first, but still developed seepage problems in the
cover. A horizontal band of dead vegetation appeared in the middle of the batter on one of
the trial plots. The die back then progressed down the batter. Excavation of the affected area
revealed an old road surface which had been used in the construction of that level of the dump.
Leachate from within the dump was moving laterally across the compacted road surface and
out into the cover. This problem showed the need to modify the internal drainage at the edge
of the benches so that leachate could be prevented from entering the cover at the toe of each
batter. Deep ripping of the old road in the trial area and re-establishment of the cover proved
effective. The other cause of dieback on these trial areas was contaminated surface runoff
from higher levels of the dump spilling onto the cover during heavy rain.
4.3 The Rehabilitation Project
Rehabilitation of the top level of the dump began late in 1989 and progressed continuously
until completion in early 1992. The chosen rehabilitation procedure for the dump involved reducing
TARLINTON 78
Management of AMD at Woodlawn Mines.
the angle of the batters to between 20% and 25%, and regrading the benches to provide controlled
drainage. The compaction of the surface was done with multiple passes of an 18 t vibrating
roller towed behind a 09 dozer. A 1 m cover of weathered rock and soil was placed over the
compacted layer and this was sown with pasture species and a variety of native shrubs and
trees.
Much greater variation than expected was revealed in the dry density of the waste material when
in situ testing was conducted during the compaction process. Samples of the material which
did not meet the specified density 2.19 t m-
3
were tested for theoretical maximum density. It
was found that in much of the dump material, it would not be possible to achieve that target
density. Monitoring of surface compaction was done from then on by a combination of field
testing and geotechnical analysis of samples. The new adopted measure of compaction was
that the material had reached 95% of its theoretical maximum density. On a few areas of the
dump, the roller was still unable to attain this density. These areas were treated conventionally,
with a 30 em clay layer placed over the waste.
Rehabilitation of the waste dump had a beneficial effect on the water balance of the site. The
runoff in the drains at the base of each level has been of suitable quality to release from site.
The completion of the waste dump project, and the construction of a small, third evaporation
dam finally allowed a net loss of water from the site in years of greater than average rainfall.
4.4 Rehabilitation Monitoring
4.4.1 Infiltration
Infiltration of rainfall into the waste rock dump was measured using a series of lysimeters placed
on the upper levels of the dump. These consisted of 200 L-stainless steel drums with open
tops, into the bottoms of which small hoses were installed. The drums were then filled with
rock and buried near the surface of the dump. The ground over the drums was then compacted
with the roller and the 1 m earth cover put in place. After several months, each lysimeter was
pumped out and the volume of water extracted was compared with the amount of rainfall tor
the same period. In most cases, the lysimeters yielded less water than would be expected had
a seal not been put in place, inferring that rainfall infiltration had been effectively minimised by
the seal.
4.4.2 Oxygen and temperature
The Australian Nuclear Science and Technology Organisation (ANSTO) installed a series of
eight monitoring facilities on the waste dump alter rehabilitation. Holes were bored through
the cover and into the waste rock to a depth of up to 9 m. Pipes were then placed in the holes
to allow the measurement of temperature and the extraction of gas samples. The results from
these installations showed quite wide variations. The general trend at almost half the sites was
a rapid reduction in oxygen concentration with depth, with 0
2
levels being mostly under 1%
within the waste rock. The other four sites produced unusual oxygen profiles, with various patterns
of 0 2 concentrations through the cover and waste. The temperature patterns were more uniform.
Temperatures at all sites increased with depth" Temperatures as high as 34 oc were observed
when the ambient air temperature was less than 10 C. The generation of heat by sulfide oxidation
within the dump is clearly indicated. Results collected over three years showed little temperature
variation at depth, and only seasonal variations near the surface.
79 TARLINTON
Management of AM D at Woodlawn Mines.
4.00 -r.---------------,
:t 3.00
=
Q)
"& 2.00
:;:
0
i:L 1.00
0.00
JUL 91 JUL 92 JUL93 JUL 94
Fig. 1. Waste rock dump leachate flow rate over time.
4.4.3 Leachate flow
response
The flow of leachate from the
toe of the dump's lowest level
is the only measurable
seepage. Weekly
measurements of the
instantaneous flow from that
area have been recorded since
July 1991, and are presented
in Fig. 1. The rehabilitation of
the upper two levels of the
dump had been completed at
the commencement of flow
monitoring. This represents
about half the dump's surface
area. As expected, there was a general decline in leachate flow as the lower levels of the dump
were rehabilitated. This trend of decreasing flow has continued since the completion of the
project. The rate of flow reduction may have been exaggerated by lower than average rainfall
during the past year. The rate of leachate flow is still quite responsive to rainfall, but the recession
time of the flow is generally shorter since the dump has been sealed. The contaminant load
of the waste dump leachate flow has also been decreasing since rehabilitation. There appears
to be a positive correlation between flow rate and metal concentration. The zinc analyses of
leachate samples taken since January 1992 are shown in Fig. 2.
The other metals analysed showed a similar pattern of decreasing concentrations as the leachate
flow abated. The pH of the seepage has not shown the same tendency, varying from 2.9 to
3.2 over the sampling period. A range of leachate analyses are shown below in Table 1.
8000
6000
',
5
4000
I
....








-- - -- -
-
c:
N
2000
0
0 0.5
1 .5 2
Flow Rate (1/s).
Fig. 2. Relationship between concentration of zinc In the leachate and flow rate.
4.4.4 Plant behaviour In the cover
One criticism of the rehabilitation design was that the establishment of plants in the cover would
be short lived, with trees dying when their roots came into contact with the waste rock below.
This has not been the case so far. The upper levels of the dump support healthy trees of up
TARLINTON
80
Table 1. Typical waste rock dump leachate analyses.
Date pH
03-Jan-92 2.9
29-Apr-93 3.1
27-Feb-95 3.0
Conductivity Cu
(mS em-') (mg L-')
2670 230
1860
2060
125
88
Zn
(mg L-')
7650
6500
4980
Management of AMD at Woodlawn Mines.
Fe Cd so.
(mg L-') (rng L-') (mg L-
1
)
350 45 39000
350 30 32800
310 24 27900
to 6 m in height. Excavations around the base of such trees show that there is little root penetration
into the compacted rock. Tree roots generally spread laterally, after reaching the waste. It is
unknown whether this is a response solely to the density of the compacted layer, or if the acidity
of the waste also has an effect.
4.5 Cover Failures
There have been some small sections of the rehabilitated dump which have shown the typical
pattern of failure after initially good vegetation growth. Patches of vegetation on the upper levels
died off after up to two years of satisfactory growth. When investigated, these seemed to be
caused by inadequate construction rather than by a flawed design. When the sites were excavated,
they were found to have insufficient depth of cover. Parts of the seal had apparently been disturbed
by machinery when spreading the cover material. Areas identified as failing due to inadequate
cover depth were excavated and sealed with clay in July 1994. These areas represent about
0.5% of the total area of the dump. Minor seepage has also occurred at the base of two of
the dump levels, apparently as a result of lateral movement of leachate across the compacted
surface of the bench and out into the cover. Ripping done on the bench was apparently insufficient
for providing the required internal drainage.
5.0 FUTURE WORK
The major AMD control issue currently facing Woodlawn is the rehabilitation of the tailings
dams. A proposal was put forward to create a permanent water cover on the third tailings
dam after it is decommissioned, thus preventing acid generation in the long term. This would
have been a most desirable option in terms of low construction costs. Detailed hydrological
investigation of the site, however, showed that the available catchment of the dam would be
insufficient for maintaining a permanent body of water under normal climatic conditions (Ryan
and Clark 1994). Investigations into the optimum design for a dry cover on the tailings dams
are continuing.
The reduction of contaminated water storage on site has been accelerated by the recent drought,
and water loss is expected to improve when tailings dam rehabilitation commences. Water
management will, however, remain a critical issue up to and beyond mine closure.
6.0 CONCLUSIONS
The history of the Woodlawn project highlights the importance of several aspects of acid mine
drainage control. Accurate assessment of the acid producing potential of waste, tailings and
construction materials needs to be carried out early in the planning stage of a project. This
allows necessary management procedures and facilities to be prepared accordingly. Predictive
81
TARLINTON
I
Management of AMD at Woodlawn M1nes.
water balance modelling is highly sensitive to climatic variations, and planning decisions made
from such calculations need to incorporate substantial safety factors. Quality control in the
construction phase of a rehabilitation project is of equal importance to good planning if a satisfactory
result is to be achieved.
7.0 REFERENCES
Hammond, R. (1981 ). The environment and its control at Woodlawn Mines. (Woodlawn Mines:
Goulburn.)
Jododex Australia Ply. Limited. (1973). Woodlawn Project: Environmental Impact Statement.
Jododex Australia Ply. Limited. (1976). Woodlawn Project: Addendum to Environmental Impact
Statement. December 1976.
Miller, S. (1993). Woodlawn Mines. Acid forming characteristics of tailings and implications
for closure. (Environmental Geochemistry International: Sydney.)
Sing line, R.N. (1992). Waste rock dump rehabilitation at Woodlawn mines: Sealing by compaction.
In 'Proceedings of National Conference on Environmental Engineering, Gold Coast,
June 1992'. (Australian Institute of Engineers: Canberra.)
Skrinde, R.T. (1979). Water management at Woodlawn Mines. (Olympic Engineering Corporation,
Seattle.)
Southern, P. (1987). Rehabilitation of metalliferous mining wastes at Woodlawn Mine, NSW.
In 'Mining Rehabilitation '87'. (Australian Mining Industry Council: ACT.)
Ryan, P., and Clarke, W. (1994). Hydrological evaluation of wet cover on the West Dam. (AGC
Woodward Clyde Ply Limited: Brisbane.)
TAR LINTON
82
Second Australian Acid Mine Drainage Workshop {Eds N J Grundon and L.C. Bell). pp. 63-90.
MANAGEMENT OF AMD AT THE PINE CREEK MINE
M.N.R. Fawcett
Pine Creek Goldfields Limited, Stuart Highway, Pine Creek, NT 0847.
ABSTRACT
Pine Creek Goldfields Ltd operates the Pine Creek goldmine, located 220 km south of Darwin
in the Northern Territory. Operations commenced mid-1985 and are scheduled to cease in
late 1995.
AMD became a serious issue during the 1990/91 wet season as mining progressed into the
unweathered primary rock zone of the Enterprise Pit. AMD was detected leaching from primary
rock waste dumps into two creek systems.
Low permeability capping of waste dumps commenced in mid-1991. The North Waste Dump
(NWD) was completed prior to the 1991192 wet-season. Since capping, monitoring has not
detected AMD escaping from the NWD. Capping of the main waste dump (South Waste Dump)
is 95% completed. This dump also has demonstrated a measurable reduction in AMD production.
New pit developments since 1992 have taken advantage of sequential mining and backfilling
of pits, avoiding the necessity to build further above-ground waste dumps.
An innovative strategy has been employed for controlling AMD within two of the final voids on
site. This strategy is to flood the pits by diverting a portion of the flow of Pine Creek into them,
submerging the majority of the exposed primary rock.
1.0 INTRODUCTION
Pine Creek is an historical gold mining town in the top-end of the Northern Territory, located
some 220 km south of Darwin via the Stuart Highway. Gold was first discovered in the region
during the construction of the Overland Telegraph in the 1870s.
Pine Creek Goldfields Limited (PCG) operations are located on two main mining leases (MLN
13 and MLN 1130) immediately adjacent to the western edge of Pine Creek township. PCG
commenced mining the main Enterprise ore body in July 1985. This orebody was the sole source
of ore until1991, when mining commenced in the smaller adjacent Czarina orebody. In early
1993, mining operations moved to the northern area of the leases where three smaller pits were
mined. These were: North Gandy's; South Gandy's; and International. Mining ceased in early
November 1994. The processing plant will continue operating until late 1995 treating stockpiled
low-grade primary ore.
Impacts of Acid Mine Drainage (AMD) and methods of limiting its production are significant
management issues at PCG operation. AMD has caused an increase in operating costs in a
number of areas. These include increased plant maintenance costs due to corrosion caused
by acidified process water, increased reagent consumption (particularly lime) as a result of pumping
AMD from the Waste Rock Sump to the Tailings Dam, and costs associated with managing
wet-season water releases from the Process Water Dam (PWD).
83 FAWCETT
IVI<..111<-l!:JUilltOI!l Ul LIVIULU a.l t'lllbl Vl\:lbiK /1JIIIIt1
Until mid 1989, mining concentrated on weathered oxidised material from the upper sections
of the Enterprise Pit. The base of oxidation was approximately 30 m to 40 m below original
ground surface. Oxide waste produced was used for the continual construction of tailings dam
walls.
Coincident with commencing to mine significant amounts of unweathered primary rock was the
construction of two waste dumps, the South Waste Dump (SWD) and the North Waste Dump
(NWD) (Fig. 1 ). Both these dumps were constructed exclusively from primary rock. The NWD
has a surface area of approximately 10 ha and contains 900,000 bank m
3
(BCM) of primary
waste. Construction of the NWD was commenced in April1990 and was completed by August
that year. Its purpose was to provide a shorter waste haul while mining in the northern section
of Enterprise Pit. The SWD's surface area is 45 ha and it contains 9 million BCM of primary
waste.
Fig. 1 Site plan of the Pine Creek Goldfields Mlneslte.
Table 1 shows the annual average rainfall at Pine Creek. AMD became an issue for this site
during the 1990/91 wet-season when leachate was detected entering both Pine Creek and
Copperfield Creek catchments (Fig. 1 ). AMD was found to be entering Pine Creek after leaching
from the NWD, and into Copperfield Creek from the SWD via the PWD spillway.
FAWCETI
84
Table 1. Average annual
rainfall at Pine Creek.
Month Rainfall
(mm)
January 269
February 244
March 208
April 47
May 7
June 3
July
August 1
September 8
October
November
December
Total
42
113
204
1147
VI t\IVI u dl l ~ l l l t . J lAtH::tK MlflH
2.0 NORTH WASTE DUMP
A notable feature for this site was the rapidity of the
AMD production process, particularly at the NWD, as
the wet-season had only just begun. AMD was detected
at CK7 (Fig. 1 ), downstream of the NWD during routine
monitoring on 11th January 1991. Only 340 mm of
that season's 1398 mm of rainfall had been received
by that stage. Samples taken at CK7 on 14th January
had a pH of 3.0 and an EC of 7960 ~ - t S em'.
Early in the 1990-91 wet season, AMD was detected
leaching from the toe of the dump. It flowed into Pine
Creek via a small tributary, and subsequently through
the town. The AMD was detected by local residents
as well as by routine mine monitoring, and led to
considerable unfavourable coverage in the media.
Short term measures were implemented to control the
AMD until the end of the wet-season. These included
dosing the creek with lime, and collecting and pumping
the leachate into Enterprise Pit.
In early 1991, AGC Woodward-Clyde was commissioned
to design a practical method of rehabilitating the North
Waste Dump, and eliminating AMD. The design required shaping the dump to a suitable profile,
then capping it with a low permeability cover of oxide waste. The purpose of the capping layer
was to reduce the amount of water and oxygen reaching the primary waste, which in turn would
control the production of AMD.
Earthworks were commenced in June 1991 and completed by October that year. The majority
of material used in the rehabilitation was taken from a cutback of the eastern wall of Enterprise
pit in progress at that time.
Prior to the placement of the capping layer, berms were constructed and shaped to control erosion
caused by rainfall runoff. Batters were covered with an 0.8 m thick layer of oxide waste and
compacted to 95% maximum dry density (MOD). The berms were covered with a 1.2 m thick
layer of oxide waste and compacted to 98% (MOD). A total of 150,000 BCM of oxide waste
was used for capping.
The capping layer was spread by graders and compacted with a vibrating smooth drum roller
drawn by a 06 dozer. A water truck was also required to maintain optimum moisture content
for compaction.
A 0.4 m layer of uncompacted oxide waste was placed over the capping layer, followed by a
thin layer of freshly stripped topsoil. The surface was then shallow ripped, fertilised and seeded
with a mix of native tree seeds and a cover crop.
The lower batter was irrigated to establish a vegetation cover prior to the start of the wet season.
Seeding of the batters was done during the 1991-92 wet season while the berms were seeded
during the 1992-93 wet season. Vegetation is now well established, with acacias in excess
of 2 m high and eucalypts up to 1 m high.
85 FAWCETI
Management of EMBED at Pine Creek Mine.
The seed species mix was altered after the first wet season to better reflect the composition
of the surrounding woodland. This was found to be necessary as the proportion of acacias
to eucalypts was unsuitable with acacias tending to dominate. The ratio of acacia seed to eucalypt
seed in the first seeding was 1 :1.16 (by weight). It has since been changed to 1 :2.43 (by weight),
significantly improving the representation of eucalypts in subsequent seeding programs.
A monitoring program was initiated to assess the performance of the rehabilitation works. The
program comprised:
(1) Burying lysimeters at three locations on the dump to check for water infiltration
through the cover;
(2) Drilling a monitoring bore through the dump and equipping it with thermocouples
to monitor temperatures within the dump, to determine if the AMD generating reaction
had ceased;
(3) Another monitoring bore was drilled down to the water table just north (downstream)
of the dump to enable monitoring of the ground water for AMD; and
(4) Regular sampling of wet season runoff from the dump to check for AMD seepage.
Since the capping was completed in late 1991, no evidence of AMD generation from the dump
has been detected (Table 2).
Table 2.
Water analysis results from monitoring site CK7.
Attribute
Date sampled
21.1.91 13.2.92 20.12.94
pH
2.9 3.9 5.1
EC (JlS em') 10,890 850 400
Mg (mg L-') 1,976 60 21
so. (mg L-'l
14,920 410 56
Mn (mg L-') 262 5.4 1.3
AI (mg L-') 649 9.5 0.26
As A (Jlg L-') 882 8 15
CdA (Jlg L-') 1,960 74 7
CuA (Jlg L') 48,800 1,200 32
FeA (Jlg L-') 53,000 250 30
PbA (Jlg L-
1
) 264 41 10
ZnA (Jlg L-') 879,000 15,000 1,400
AMeasured in filtrate
The cost of capping the NWD was $36,000 ha _,_ This did not include the cost of mining and
hauling the 150,000 BCM of oxide waste used for capping, as it was necessary to mine this
material as part of a cut back of the eastern wall of Enterprise Pit. The design for capping and
rehabilitation for the South Waste Dump was based on the lessons learned from this project.
FAWCETI 86
Management of AM D at Pine Creek Mine.
3.0 SOUTH WASTE DUMP
The South Waste Dump is located approximately 1 km south-west of the Enterprise Pit, and
immediately north of the tailings dam. It has a surface area of approximately 45 ha, and was
the principal waste dump for the majority of Enterprise and Czarina primary waste. Waste
characterisation studies were performed by AGC Woodward-Clyde in 1991 on a number of samples
of primary waste rock representing the majority of the material dumped at the South Waste Dump.
The waste rock in the dump is divided into three geological units: (a) Spotted Silt Horizon (SSH);
(b) Upper Mine Greywacke (UMGW); and (c) Lower Mine Greywacke (LMGW).
Results of the test-work indicated that both the SSH and the UMGW ranged from moderately
to non-acid forming, and in some cases had a potential net acid neutralising capacity. However,
test-work on the LMGW indicated that the majority of samples were moderately to strongly acid
forming.
The South Waste Dump started generating significant amounts of AMD in 1990, within a year
of commencing the primary waste dumping. AMD from the South Waste Dump has had a
detrimental effect on water quality in the Process Water Dam (PWD). In the 1991 dry season,
the pH dropped to as low as 2.5, with a correspondingly high conductivity reading, and high
levels of dissolved metals.
During an average wet season, the PWD overflows into Copperfield Creek via a spillway channel
at its southern end. A consequence of the poor water quality in the PWD has been difficulty
in managing the quality of wet season water releases. As it is not possible to control the quantity
of water discharged, it was therefore essential that the quality be improved.
A short term solution to improving PWD water quality has been to construct a sump at the
downstream toe of the SWD to collect leachate that is then pumped up to the Tailings Dam.
In the long term, the only solution was to cap the dump in a similar manner to the NWD. A
major cost impact of pumping acidic water to the Tailings Dam has been to increase process
plant lime consumption from an average of 1.8 kg r' up to 3.5 kg r' over the last three years.
Capping of the South Waste Dump commenced in late 1991, and continued as suitable capping
material became available. By March 1995, less than 3 ha remained to be capped.
The first 7 ha of dump capping utilised spent heap leach materiaL This material had been crushed
to< 25 mm, and agglomerated prior to stacking and leaching. It has not proved to be a successful
capping medium due to its poor erosion resistance and failure to adequately support plant growth
compared with run of mine oxide waste rock. With the commissioning of Czarina Pit in July
1992, run of mine oxide waste rock became available, and it has been used for capping the
remainder of the dump.
As an erosion control measure, contoured drainage banks have been constructed over the majority
of the dump.
Capping costs have been significantly reduced each year since work on the SWD commenced.
The first 7 ha of the SWD cost $27,000 ha-', but by 1993, this had been reduced to less than
$7000 ha'. The cost reductions were achieved initially by refining the spreading and compaction
methods, and ultimately, by integrating the capping process into the mine waste dumping operations.
These costs clearly demonstrate that, wherever possible, waste dump construction must be
87
FAWCETI
Management of EMBED at Pine Creek Mine.
thoroughly planned, and that dump rehabilitation is an integral, progressive and scheduled part
of the mining operation; not left as an after mining clean-up job.
The results in Table 3 demonstrate that since capping began in 1991 in conjunction with a
catchment management plan, there has been a significant reduction in the amount of AMD reaching
the PWD. Further improvements are expected when capping is completed in late 1995.
Table 3. Water analysis results from the process water dam.
Attribute Date sampled
23.4.91 18.3.92 3.2.95
---
pH 3.6 4.2 4.5
EC (JLS em-') 2,400 1,700 1,200
Mg (mg L-') 240 110 51
S0
4
(mg L-') 1,800 1,000 530
Mn (mg L-
1
) 27 12 5.7
AI (mg L-') 70 19 3.7
AsA (J.lg L-') 12 10 20
CdA (JLg L-') 300 10 19
CuA (JLg L-') 4,700 3,500 680
FeA (JLg L-') 4,800 3,600 280
PbA (J.lg L-
1
) 280 10 17
ZnA (JLg L-') 102,000 30,000 6,900
AMeasured in filtrate
4.0 GANDY'S HILL PITS
During 1993 and 1994, three small pits were developed north of the original Enterprise Pit.
They are: South Gandy's; North Gandy's; and Inter-national. Mining was completed in these
pits by November 1994. The development of these pits took advantage of the opportunity to
integrate waste management into the mining plan, and this lead to significant savings in waste
handling and rehabilitation costs.
For the purposes of waste management, all primary waste produced from these pits was treated
as potentially acid forming, and required capping.
Careful scheduling of mining operations allowed !or the sequential backfilling of voids, with primary
waste placed in the lower sections followed, by capping layers of oxide waste. It was only
necessary to double handle a small amount of oxide waste to complete capping once mining
ceased.
The one remaining void at the southern end of South Gandy's Pit was flooded by diverting a
portion of Pine Creek's flow into it in a similar manner to Enterprise Pit.
FAWCEH
88
Management of AM D at P1ne Creek Mine.
5.0 ENTERPRISE PIT
An innovative rehabilitation technique has been employed to deal with the potential for AMD
production in the Enterprise Pit final void.
Mining of the Enterprise Pit commenced in mid 1985 and was completed in January 1993. During
that period, 12.3 million m
3
of ore and waste were extracted. The dimensions of the final void
are approximately 900 m long, 300 m wide, and 140 m deep (RL 1 060).
In March 1993, ERA Environmental Services (ERA) was commissioned to prepare a post-<Jperational
plan of management for the Enterprise pit based on hydrology and geochemistry considerations.
ERA prepared this plan with the assistance of the CSIRO, PAWA (Power and Water Authority}
and Centre for Water Research (WA}.
The Plan of Management recommended that the best possible water quality in Enterprise Pit
would be achieved by rapidly flooding it with water diverted from Pine Creek, at the northern
end of the pit. It has taken two wet-seasons to fill the pit to RL 1198, which is some 10 m to
15m below most of the pit rim.
If allowed to fill naturally by ground water inflow and rainfall, the pit would not overflow, but would
reach a steady state level at approximately RL 1170 - 1180 over an extended period of time.
The water quality in this case was likely to be poor, as there will be an annual production of
acid and metals from the exposed pit walls. If this occurred, there is a substantial risk that the
pit may end up in a similar condition to some of the other decommissioned sulfide pits in the
Top End, where the water is of very poor quality and strongly acidic.
Flooding Enterprise Pit required the construction of a diversion weir in Pine Creek and a channel
to connect it to the northern end of the pit. The weir and channel divert the majority of the wet
season flow of Pine Creek upstream of the mine into the pit.
Government approvals for this proposal were sought and received in late 1993. Construction
of the diversion structures was completed in December 1993, and the first water flowed into
the pit on the 23rd December 1 993. It filled to its design level (RL 1198) by early March 1995
and now contains 6,800 ML of water.
Additional benefits of flooding the pit were:
(1} Fulfilment of the local community's expectations that there will be a large lake
on the edge of town, with possible recreational uses;
(2} A reduced risk to public safety through the reduction in exposed height of pit walls;
and
(3} To control the ground water level so as to prevent any adverse impact on the ground
water supply to Pine Creek.
Table 4 shows water analyses from the Enterprise Pit.
89 FAWCETT
Management of EMBED at Pine Creek Mine.
Table 4. Water analysis results from the Enterprise Pit
Attribute
pH
EC (!lS em')
Mg (mg L')
S0
4
(mg L')
Mn (mg L')
AI (mg L')
AsA (!lg L')
CdA (!lg L"')
CuA (!lg L"')
FeA (!lg L')
PbA (!lg L-')
ZnA (!19 L"')
AMeasured in filtrate
6.0 CONCLUSIONS
6.3.91
6.5
1,490
107
800
13.1
0.1
22
39
<40
0.9
11,400
Date sampled
6.8.93
7.1
1,300
70
660
3.7
26
47
19
28
8.3
3
6,900
21.3.94
6.1
220
14
77
0.7
0.05
4
2
13
84
38
1,100
The application of a diverse range of approaches has been required to control AMD at this site.
The situation that arose at the North and South Waste Dumps would not have been as severe
or as expensive to rectify, if the following points had been adhered to:
(1) Waste characterisation must be undertaken before operations commence and
the results checked as mining progresses;
(2) The scheduling of waste placement must be given the same priority as scheduling
ore production, and a waste dump plan developed at the same time as the mine
plan;
(3) Mining crews, supervisors and planners must realise that long term waste handling
strategies are an integral part of their job;
(4) Site decommissioning planning should begin at the project design stage, and then
regularly reviewed throughout the operations life;
(5) Catchment management strategies should be integrated into the project design
stage;
(6) Wherever possible rehabilitation should be progressive- not left until mining has
finished; and
(7) Environmental approval documents must be referred to frequently and not be
considered to have served their purpose once a project is approved.
All of the above points were integrated into the development of North Gandy's, South Gandy's
and International Pits. This meant that the entire project area covered by these pits, and their
associated infrastructure, was ready for revegetation within six weeks of mining ceasing.
FAWCETI 90
SEcond Australian Acid Mine Drainage Workshop (Eds N.J Grunden and L C. Bell). pp. 91-100
DEVELOPMENT OF A WASTE DUMP MANAGEMENT
STRATEGY AT MT. LEYSHON GOLD MINE
M.S. Orr
Normandy Poseidon Group, P.O. Box 6080, Townsville Mail Centre, Old 4810.
ABSTRACT
Much of the waste rock at Mt. Leyshon Gold Mine is sulfidic and has the potential to generate
acid mine drainage. A strategy has been developed at the mine to minimise the risk of acid
mine drainage, incorporating waste characterisation, selective waste handling, and sealing of
the bulk waste using compacted rock cover. An innovative method of compaction was developed
for this purpose. The performance of the seal system has been assessed via the modelling
of long-term infiltration rates, and via direct measurement of oxygen and temperature gradients
within the dumps. An ongoing quality control program has been implemented to monitor compliance
with specifications.
1.0 INTRODUCTION
Background information on MI. Leyshon Gold Mine (MLGM) has been presented previously
in these proceedings (Orr 1995). As commonly found at many metalliferous mines, the primary
(unoxidised) waste rock at MLGM contains varying levels of sulfides, mainly in the form of pyrite.
When exposed to oxygen and water, these minerals will oxidise to produce sulfuric acid. Sufficient
water passing through the waste will result in metal-rich seepage exiting the dump, commonly
known as acid mine drainage (AMD). The rehabilitation strategy for the waste dumps must
therefore be designed to minimised the potential for AMD. This paper describes the development
of a rehabilitation strategy to meet this objective.
2.0 REHABILITATION OPTIONS FOR THE CONTROL OF ACID MINE DRAINAGE
The current knowledge of this issue has defined two general strategies for the control of AMD,
with emphasis on the roles played by oxygen and water. As oxygen availability is a limiting
factor to sulfide oxidation rates within a waste dump, restricting oxygen entry into the dump
can reduce the rate of acid generation (Bennett et aL 1992). Alternatively, if water flux into the
dump can be reduced. the volume of acid leachate exiting the dump will be less, reducing the
requirements for containment. Mechanisms have been proposed by which this reduction in
flux through the dump could also result in a reduction in total metal load leaving the dump.
For example, the channelling of flow along preferential pathways within the waste body (Eriksson
and Destouni 1994).
At MLGM, mine management has considered both these strategies. Water infiltration into a
dump can be reduced to very low levels by constructing a low permeability compacted cover
over the bulk waste. A compacted cover can also be an efficient means of preventing convective
transport of oxygen into a dump (Bennett et al. 1992; Bennett and Pantelis 1991 ). The ability
of such a layer to restrict oxygen diffusion may be significantly enhanced if some of the pore
space within the seal is occupied by water, through which oxygen has a very low diffusion rate.
In the climate experienced at MLGM, with an extended dry season each year, it is difficult to
maintain complete saturation within a compacted seal. However, modelling has indicated that
at least partial saturation can be achieved for extended periods.
91
ORR
Wasta Dump Management at Mt Leyshon Gold Mine.
3.0
DEVELOPMENT OF A REHABILITATION STRATEGY FOR MLGM
3.1 Definition of the Rehabilitation Strategy
The chosen rehabilitation strategy for the dumps at MLGM is the construction of a sealing layer
over the bulk waste that will:
(1) Reduce infiltration to very low levels;
(2) Prevent convective transport of oxygen; and
(3) Have the potential to reduce diffusive transport of oxygen.
This strategy is shown schematically in Fig. 1.
M a x i m i ~
runoff
Minimise
seepage
Native shrubs and
pasture species
Rainfall
Minimise o
2
entry
SJ!fidic Waste
Fig. 1. The chosen waste dump strategy for Mt Leyshon
Gold Mine.
3.2 Development of
the Strategy
3.2.1 Compaction trials
A key issue in implement-
ing this strategy was the
availability of material suit-
able for seal construction.
The clay reserves on-site
are limited and the costs
involved in using clay
would be relatively high.
Oxide heap-leach material
has been used to seal
some areas of the dump,
but the saturated hydraulic
conductivities achievable
through compaction of this
material are limited to
about 10-4; m s-
1
A paper
by Singline (1992)
describes the compaction
of in situ waste dump material to form a low permeability seal at Woodlawn Mine. This approach
was seen as having potential benefits at MLGM, and a number of compaction trials were conducted
on-site to explore this possibility. Shown in Fig. 2 is a plan of the first compaction trial using
mine waste. The four major waste types at the mine were laid in strips to a depth of about 1 m.
Three different compaction methods were run at right angles to these strips, namely:
ORR
( 1) A control strip with no compaction except for that produced by the bulldozer when
pushing out the strip;
(2) A fully loaded haul truck with a total weight of about 360 t; and
(3) A BH-1300 impact roller (Plate 1 ). When towed at a speed of around 12 km h .. ,,
the large steel (8 t) drum of this machine impacts on the ground with great force.
The claimed advantages of this device over the more usual vibrating drum roller
are good compaction at depth and relative insensitivity to the moisture content
of the material.
92
20m 20m
10m
10m
10m
Waste Dump Management at Mt. Leyshon Gold Mine
20m 20m
Granite-
dolerite
Control
(no compaction)
~
Fully loaded
truck (360 t approx.)
-----------
-+--
Impact roller
Fig. 2. Layout of the waste rock compaction trial at Mt Leyshon Gold Mine.
Plate 1. The BH-1300 Impact roller compacting porphyry seal at Mt Leyshon
Gold Mine.
Testwork conducted during the trial included compaction tests (sand replacement method) and
hydraulic conductivity. Hydraulic conductivity was measured directly using the device shown
in Fig. 3. This 'infiltrometer' was built on .. site and measures infiltration rate over a relatively
large area (1 m
2
).
The key finding from the first compaction trial was that the porphyry waste was soft enough
to be readily compacted (none of the other waste types were shown to be suitable for use as
a seal material). Both the haul truck and the impact roller were able to obtain high densities
93 ORR
Waste Dump Management at Mt Layshon Gold Mina,
IMNOI<"'
"\.,'JL__
lnrutromcter
Expanding Foam
.:,-.:.:;t
Wetted B.Jib
- - - - - - - - - - - - - ~ " '
I
I
Expanding Foam
Fig. 3. The lnflltrometer design In use at Mt Leyshon Gold Mine.
at depth (Fig, 4), On
both the truck and
impact roller com-
pacted surfaces,
hydraulic conductivi-
ties of 1 o-
7
m s-' to
1 o-a m s-' were mea-
sured, an encourag-
ing result. The use
of a truck was discon-
tinued in further test-
work due to the good
apparent performance
of the impact roller
(the strategy of ob-
taining compaction
through routine truck
travel was considered
unrealistic if a close
specification was
105 r - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ~
~ 100
0 ~
i 95

0
1' 90
~
1i5 85
*


'
'

80 L_ ______ J_ ______ _L ______ _J ______
200 400
Depth (mm)
600 BOO
Impact roller (20 passes) *Loaded truck (20 passes)
Control
Fig. 4. Compaction results for porphyry In the first
compaction trial (each point a mean of four
replications).
required).
It was thought that the perform-
ance of the impact roller in the
original trial may have been com-
promised by the short run-up
available, resulting in average
speeds lower than the optimum
of 12 km h-'. A larger trial was
conducted on a flat area of dump
approximately 50 m by 50 m, using
a 1 m thick layer of porphyry
waste, The trial gave higher
densities at the surface, with an
average standard compaction of
103% (Fig. 5), Equally importantly,
compaction declined little with
depth to at least 0,5 m. Also
shown in Fig. 5 are the results
obtained on porphyry using a 16 t
vibrating sheep's-foot roller, typical
of machines often used in this
application, By comparison with
the impact roller, compaction declined very quickly with depth (fewer passes were used with
the vibrating roller as this machine travels much more slowly than the impact roller; it was
considered that a greater number of passes would not be a practical option for routine rehabilitation
work),
3,2.2, Modelling of long-term performance
To quantify the performance of the possible porphyry seal configurations, infiltration was modelled
using the 'SWIM' modeL The role of mathematical modelling in the assessment of rehabilitation
ORR 94
Wasta Dump Management at Mt Layshon Gold Mine.
110
~
c: 105
.Q
1:J
"'
a.
E 100
8
"0
(ij
"0 95
c:
<7l
90
'
0


100 200 300 400 500
Impact roller {12 passes)
16 t vibrating roller (6 passes)
Fig. 5. Results from additional compaction trials on porphyry
waste (Each point Is the mean of 4 replicates).
strategies is very
important, because of
the cost and time
required for field trials,
The very variable rainfall
received on-site means
that meaningful
estimation of long-term
infiltration rates via
direct measurement
would require several
years of monitoring at
least. The infiltration
modelling methodology
applied at MLGM is
described in Crees et al.
(1994). A key feature
is the use of 26 years
of pluviometer data in
the modeL The results of the modelling (Table 1) indicate that very low levels of infiltration can
be achieved using the impact roller on the porphyry, and that benefits are obtained from increasing
seal thickness to 1.0 m, and from decreasing saturated hydraulic conductivity to 1 0-a m s-'.
Table 1. Results of SWIM modelling of the porphyry seal, showing average annual
Infiltration through seal {as water depth and percentage of total precipitation)
for 26 years of pluvlometer data.
Saturated hydraulic
conductivity
(m s-')
1 x 1 o-
7
5 X 10-"
100 em
24 mm (3.9%)
11 mm (1.7%)
3.2.3 Geochemical assessment of the seal
Seal thickness
50 em
35 mm (5.8%)
18 mm (3.0%)
Having examined the physical characteristics of the porphyry material, it was necessary to consider
its geochemical characteristics. Representative samples were collected from drill core and areas
of compacted seal, and analysed for sulfur and net acid producing potential (NAPP). The results
showed that, unlike the other main primary wastes, the porphyry has significant acid neutralising
capacity. On the basis of these samples, porphyry with less than about 0.8% sulfur could be
expected to have negative NAPP (Fig. 6).
More intensive sampling of drill cores were then conducted to provide a better understanding
of sulfur distribution within the waste body. A three-dimensional computer model of the porphyry
waste was constructed, using normal grade-control methodology. This allowed the production
of the tonnage-sulfur distribution shown in Fig. 7. This information confirmed that sufficient low-
sulfur porphyry would be available to cover the entire waste dump areas to a depth greater than
1m.
95
ORR
Wasta Dump Management at Mt Layshon Gold Mina.
<?
100
C5
(J)
N
I
6
50
~
c
~ ,.0
2
0
c.
0 -----
""

c

'[5
:J
0 0
1:l
-50 ~ "'. e
c.
'0
'[5
co
~ -100
0




Rancaa sarrplas
o I o RaniWtu seal
1 2
Sulphur content(%)
Fig. 6. Relationship between sulphur concentration and net
acid producing potential of the porphyry waste.
"'6
c
,g
:_
"'
" 4

"
12
:J
()
0 ~ - - ~ . L - - ~ - - ~ - - - - ~ - - ~ - - ~ - - - - ~ - - - ~ - - ~
0 0,1 02 0,3 0.4 0,5 0.6 0.7 0.8 0.9
Sulphur content (%)
Fig. 7. Sulphur distribution of porphyry waste.
4.0 MATERIAL AVAILABILITY AND SCHEDULING
As described in the preceding section, field compaction trials, infiltration modelling and waste
characterisation confirmed the feasibility of using porphyry as a waste dump sealing materiaL
It was then necessary to consider the availability of rehabilitation material over the life of the
mine, both with regard to total availability and timing of material from the pit.
ORR 96
Wasta Dump Management at Mt Layshon Gold Mine.
110
"" c 105
~
c.
E too
8
'E
"' u 95
c
~
(f)
'


90 ~ - - - - ~ - - - - ~ - - - - - - ~ - - - - ~ - - - - - J
0 100 200 300 400 500
Impact roller (12 passes)
16 t vibrating roller(6 passes)
Fig. 5. Results from additional compaction trials on porphyry
waste (Each point Is the mean of 4 replicates).
strategies is very
important, because of
the cost and time
required for field trials.
The very variable rainfall
received on-site means
that meaningful
estimation of long-term
infiltration rates via
direct measurement
would require several
years of monitoring at
least. The infiltration
modelling methodology
applied at MLGM is
described in Crees et al.
(1994). A key feature
is the use of 26 years
of pluviometer data in
the modeL The results of the modelling (Table 1) indicate that very low levels of infiltration can
be achieved using the impact roller on the porphyry, and that benefits are obtained from increasing
seal thickness to 1.0 m, and from decreasing saturated hydraulic conductivity to 1 0_. m s-'.
Table 1. Results of SWIM modelling of the porphyry seal, showing average annual
Infiltration through seal (as water depth and percentage of total precipitation)
for 26 years of pluvlometer data.
Saturated hydraulic
conductivity
(m s-')
100 em
1 X 1 Q-
7
24 mm (3.9%)
5 x 10_. 11 mm (L?%)
3.2.3 Geochemical assessment of the seal
Seal thickness
50 em
35 mm (5.8%)
18 mm (3.0%)
Having examined the physical characteristics of the porphyry material, it was necessary to consider
its geochemical characteristics. Representative samples were collected from drill core and areas
of compacted seal, and analysed for sulfur and net acid producing potential (NAPP). The results
showed that, unlike the other main primary wastes, the porphyry has significant acid neutralising
capacity. On the basis of these samples, porphyry with less than about 0.8% sulfur could be
expected to have negative NAPP (Fig. 6).
More intensive sampling of drill cores were then conducted to provide a better understanding
of sulfur distribution within the waste body. A three-dimensional computer model of the porphyry
waste was constructed, using normal grade-control methodology. This allowed the production
of the tonnage-sulfur distribution shown in Fig. 7. This information confirmed that sufficient low-
sulfur porphyry would be available to cover the entire waste dump areas to a depth greater than
1 m.
95
ORR
Waste Uump ManagemGnt at Mt L9yshon Gold Mina.
2
0
(fJ
N
I
Dl
c.
(ii
""
c
2
0
Q.
Dl
c
'(j
:::>
1J
e
Q.
1J
'(j
"'
100
50
0 ---

0
-50
0



"'.0



0
I
Rancaa sarrples

o'
0
o Fitm hsitu seal

-100
0 2
Sulphur content(%)
Fig. 6. Relationship between sulphur concentration and net
acid producing potential of the porphyry waste.
-;o6
c

].
0

4
.2
L
" u

o 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Sulphur content (%)
Fig. 7. Sulphur distribution of porphyry waste.
0.8 0.9
4.0 MATERIAL AVAILABILITY AND SCHEDULING
As described in the preceding section, field compaction trials, infiltration modelling and waste
characterisation confirmed the feasibility of using porphyry as a waste dump sealing material.
It was then necessary to consider the availability of rehabilitation material over the life of the
mine, both with regard to total availability and timing of material from the pit.
ORR
96
Waste Dump Management at MI. Leyshon Gold Mine.
Figure 8 shows a rehabilitation materials balance for the mine. Such planning confirms the
availability of sufficient materials to support the proposed rehabilitation strategies for the different
sites. In this example, sufficient low-sulfur porphyry is available to cap all the waste dump and
one of the three tailings dams on-site (clay being scheduled for the other two dams). Topsoil
supply is also considered. A number of different budgets were prepared to consider different
strategies.
Fig. 8. Rehabilitation materials balance for the site.
With validation of the strategy with regard to total supply, attention was given to materials
scheduling. The production of porphyry water from the pit is controlled by the mining schedule.
Stockpiling and rehandling must therefore be carried out to provide a supply for rehabilitation
purposes as required. However, the degree of rehandling can be minimised with careful planning,
as shown in Table 2.
5.0 ADDITIONAL CONSTRUCTION DETAILS
5.1 Selective Handling of Bulk Waste
Before application of the porphyry seal, oxide waste and low-sulfur primary waste is placed around
the periphery of the dump to provide additional isolation of the more reactive material (Fig. 9).
The porphyry is dumped over this surface before shaping and compaction. Under the conditions
of the mining contract at MLGM, the cost of this selective waste handling and dump shaping
is borne by the mining contractor, a mechanism that encourages efficient waste handling.
5.2. Compaction on Dump Batters
Testwork showed that the impact roller did not achieve good compaction running down the batters,
as it tended to push the porphyry down with it. In addition, the average 1 :4 slope was too steep
to allow safe compaction along the contour. These difficulties were avoided by adopting a terraced
landform (Fig. 10) which reduces the effective slope angle from 1:4 to 1:5, allowing the impact
roller to travel along the slope at optimum speed. Each compacted bench overlaps the preceding
97 ORR
Waste Dump Management at Mt Leyshon Gold Mine.
Table 2. Life-of-mine rehabilitation materials scheduling.
Materials and operation Tonnage of materials in each year Total
1994 1995 1996 1997 1998 1999 2000
Porphyry
Northern waste dump 158 159 0 0 0 0 0 317
Southern waste dump 65 65 0 0 0 0 0 130
Eastern waste dump 633 633 633 633 634 634 305 4105
Scats dump 0 0 0 0 0 206 206 412
Northern tailings dam #1 0 0 0 578 578 0 0 1156
Total required 856 857 633 1211 1212 840 511 6120
Porphyry pit production 2823 1481 1028 564 246 476 382 7000
Yearly excess porphyry 1967 624 395 0 0 0 0 2986
Porphyry rehandle 0 0 0 647 966 364 129 2106
Porphyry stockpile 1967 2591 2986 2339 1373 1009 880
Clay
Southern tailings dam 0 0 0 0 0 0 452 452
Northern tailings dam #2 0 0 0 0 0 0 738 738
Total required 0 0 0 0 0 0 1190 1190
Clay production 0 0 0 0 0 0 1190 1190
Yearly excess clay 0 0 0 0 0 0 0 0
Clay rehandle 0 0 0 0 0 0 0 0
Clay stockpile 0 0 0 0 0 0 0 0
bench, resulting in a zone of continuous compaction from the base to the top of the dump.
The face of each porphyry bench is rock armoured to minimise erosion, and the lip of each bench
is shaped to provide a graded drain. This terraced landform has the additional benefit of reducing
the erosion hazard (less steep slopes and shorter effective slope lengths), and allowing easier
access for seed and fertiliser application, maintenance, etc.
6.0 QUALITY CONTROL
Following the implementation of the waste dump sealing strategy, a quality control program
was implemented. This has two functions. Firstly, it allows a rapid assessment of quality of
work, so that practices can be modified in response to changes in the results being obtained.
Secondly, it provides an ongoing record of results to allow verification that commitments made
in the EMOS (the key approval document) are being fulfilled. All rehabilitated areas are sampled
on a 50 m x 50 m grid. At each point, the following is recorded:
(1) Seal thickness;
(2) Seal surface compaction;
(3) Seal sulfur content and NAPP; and
(4) Topsoil thickness.
The data is recorded with survey coordinates, allowing the production at any time of a plan of
the dump with all these parameters shown.
ORR
98
Waste Dump Management at Mt. Leyshon Gold Mine.


Fig. 9. Formation of dump batters.



/
Contour bank
r
" .. " .........

\ \. '-. / "' " ... " " " Average l!llope "'1:4
\ I I },_-' _: .': .': ._ -----------.
--' / ::>-... \, / .:> .. .. .. .. ..
1\\.- 1\\. 1\\ .. !...........
,, I,, I,, I..AAAA6A446A.6.4
"/ "/ "/--. .......... .. \ --
\ I \ I I I I I \ I I - I I II. '' ":
'/ // ' / // \. / /> ' / // ' / <-"-,-"....._"',....._"......_"'--"'-"-"." .. ",.",."""
-- \ .......... -\ ......... -- \ .......__ -- \ ' -- \ ' ...._ \ "-... .;:::: -\--...
1\\ / I I \ / I /\\ / I I \ / 1\\ / I I \ / I 1\\ /
,, \,, ,, \,,1 '' \,, ,,
Fig. 10. Detail of compaction on dump batters.
7.0 COSTS
The impact roller and tractor (second-hand) cost about $200,000. It is estimated that the operational
cost of constructing the porphyry seal (1 m thickness) is $4100 ha-'. This does not include
the cost of selectively handling the porphyry or shaping of the dump, as these activities are
carried out as an integral part of mining operations, and are the responsibility of the mining
contractor rather than MLGM. The total cost of dump rehabilitation for the flat surfaces (includino
sealing, soil replacement, erosion control structures and quality control) is about $12,000 ha-'.
99 ORR
Waste Dump Management at Mt Leyshon Gold Mine
8.0 REFERENCES
Bennett J.W., Harries, J.R., and Ritchie, A.I.M. (1992). Chemical activity and water balance
of the overburden heaps. In 'Rum Jungle Rehabilitation Project Monitoring Report 1986-88'.
{Eds. M. Kraatz and R.J. Appplegate.) pp. 103-119. (Conservation Commission of the
Northern Territory: Darwin.)
Bennett J.W., and Pantel is G. (1991 ). Construction of a waste rock dump to minimise acid mine
drainage- A case study. Proceedings of the Second International Conference on the
Abatement of Acidic Drainage, Montreal, September 16-18, 1991. pp. 299-318.
(CANMET: Montreal).
Crees M.R., Smetterm K., and Orr M. (1994). Evaluation of the potential of artificial soil covers
to reduce infiltration. Proceedings of the 'Water Down Under 94' Conference, Adelaide,
21-25 November 1994. (International Association of Hydrogeologists and the Institution
of Engineers Australia).
Eriksson N., and Destouni G. (1994). Modelling field-scale transport of weathering products
in mining rock dumps. Proceedings of the International Land Reclamation and Mine
Drainage Conference and Third International Conference on the Abatement of Acidic
Drainage, Pittsburgh, April24-29. pp. 50-59. (United States Department of the Interior).
Orr, M.S. (1995). Control of acid mine drainage through water management at Mt. Leyshon
Gold Mine. In 'Proceedings of the Second Australian Acid Mine Drainage Workshop',
Charters Towers, Queensland. 28-31 March 1995. {Eds. N.J. Grundon and L.C. Bell).
pp. 67-73. (Australian Centre lor Minesite Rehabilitation Research: Brisbane, Australia).
Sing line, R.N. (1992). Waste Rock Dump rehabilitation at Woodlawn Mines: sealing by compaction.
ORR
Proceedings of the National conference on Environmental Engineering. Gold Coast,
Queensland, June 17-19. pp. 199-202.
100
Second Australian Ac1d M1ne Dramage Workshop {Eds N J Grundon and L C Bell). pp 1 0 1 ~ 116
APPLICATION OF OXIDATION RATES
IN REHABILITATION DESIGN
A.I.M. Ritchie
Environmental Science Program, Australian Nuclear Science and Technology Organisation,
Private Mail Bag 1, Menai, NSW 2234.
ABSTRACT
While acid base accounting is a guide to whether drainage from pyritic waste may be acid or
not, pollutant loads and pollutant concentrations in drainage depend on the rate of pollutant
generation in the waste. A major determinant for pollutant generation rates is the intrinsic oxidation
rate (lOR) of the waste material. Some typical measured values of the /OR are given and simple
expressions presented to indicate how the lOR determines some timesca/es of importance.
A simple mathematical model is presented that allows 'back of envelope', estimates to be made
of pollutant loads and timescales for pollutant generation. A case study is presented where
measurements of the lOR have been used to assess the practicality and efficacy of management
options.
1.0 INTRODUCTION
Acid base accounting has been used lor a long time to assess whether or not drainage from
pyritic wastes (mainly tailings dams, waste rock dumps, and coal reject piles) could be acid.
There has been considerable discussion on the most appropriate methods to use in acid base
accounting and what the answers mean in terms of pyritic waste management (see for example,
Miller 1992). Acid base accounting provides no information on the rate at which pyritic material
oxidises. It is the rate which determines pollutant loads and pollutant concentrations in drainage
and hence the environmental impact of drainage from the pyritic wastes.
Measurements in the waste rock dumps of the Rum Jungle mine site were used to evaluate
the pyrite oxidation rate at various points (Harries and Ritchie 1981) and to show that the oxidation
rate depended very much on oxygen transport rates to the oxidation sites, the dominant ones
at Rum Jungle being diffusion and convection {Harries and Ritchie 1985). A series of mathematical
models of increasing sophistication (Ritchie 1977; Davies and Ritchie 1986, 1987; Pantelis and
Ritchie 1992; Pantelis 1993) demonstrated the interaction of the intrinsic oxidation rate, gas
transport and heat transport in defining the oxidation rate at a particular part of the dump. It
became apparent that the high temperatures (in some cases about 55 C) found at Rum Jungle
were not a prerequisite of a dump generating pollutants at an environmentally significant rate.
This was borne out by measurements in waste rock dumps at a number of mine sites. A good
example is a waste r9ck dump at the Aitik mine site in Arctic, Sweden, where pyrite is oxidising
at temperatures close to 0 oc (Bennett et al. 1994).
Measurements of other bulk parameters which impact on the overall oxidation rate, thermal
conductivity (Blackford and Harries 1985), air permeability {Bennett 1989) and the oxygen diffusion
coefficient, have been carried out in a number of waste rock dumps. The current assessment
is that from the point of view of assessing management options, these parameters do not vary
substantially in a waste rock dump or from waste rock dump to waste rock dump (Ritchie 1994a).
101 RITCHIE
~ *
Oxtdation Rates and Rehabilitation Design.
Section 2 contains a discussion on the physical interpretation of the lOR and how it can be
used to assess timescales. Section 3 discusses the range of the lOR found in field measurements.
Section 4 presents a simple model which allows straightforward evaluation of overall oxidation
rates to be made for large dumps where advective gas transport can be neglected, and Sections
5, 6 and 7 discuss a field example where the lOR has been used to assess the efficacy of various
broad management options. Section 8 contains a brief description of how oxidation rates can
be used to assess the effectiveness of a rehabilitation scheme.
2.0 THE INTRINSIC OXIDATION RATE (lOR)
2.1 Definition and Use
The lOR is simply the rate of consumption of oxygen by material in the waste under conditions
which apply to that material. The most practical unit to use is kg (0
2
) m-
3
s-', since field
measurements produce a result in this form, and this is the most convenient form to use in
calculating pollutant production rates.
The unit, kg (0
2
) kg-' (waste) s-' may, in many cases such as in comparing laboratory
measurements, be appropriate since it removes the density dependence of the material examined.
In principle, the lOR is a function of a large number of parameters. Typical examples are: physical
conditions (e.g. temperature, particle size distribution, physical form of pyrite); chemical conditions
(e.g. eh, pH, Fe
3
/Fe' ratios); and on bacterial conditions (e.g. bacterial ecology, bacterial
numbers). In practice, the inclusion of all of these variables leads to very complex and unwieldy
mathematical models. The complexity is almost certainly not warranted given the inhomogenity
of wastes, particularly waste rock dumps. It is technically very difficult to measure many of these
parameters in the field. Finally, data on the detailed dependence is far from complete even
in laboratory experiments. Moreover the field measurements we have carried out in waste rock
dumps indicate that the lOR tends to be characteristic of large regions, metres to tens of metres
in extent. At the current of stage of our understanding of pollutant generation rates in waste
rock dumps, it is important to know the magnitude of the lOR, and how the magnitude impacts
on the overall pollutant generation rate in a waste rock dump. It is sensible to include dependence
on different variables if and when it becomes obvious that such inclusion is important from the
point of view of managing AMD. In this paper the magnitude of the lOR quoted from field
measurements is that from measurements within a waste rock dump. That quoted from modelling
studies is the maximum value the lOR can take. Modelling studies, for example, using FIDHELM
(Pantelis 1993) can include dependence on oxygen concentration, pyrite concentration and
temperature.
2.2 Relevance of lOR to Management of AMD
In assessing management options for AMD at either existing or proposed mine sites, it is useful
to be able to quantify some timescales. The time between dumping waste rock and the time
for AMD to appear and the time for all the sulfide material to be used up are two such examples.
Similarly it is useful to assess the magnitude of the pollution load and drainage. The lOR can
be used to provide estimates of all of these.
The timescale to establish the acid conditions which put metal pollutants, such as copper, into
solution is determined by the time for dissolution of the fast-reacting, acid-consuming material
in the wastes. Carbonates usually form the bulk of such material.
RITCHIE 102
The timescale is
1 - , P,
' ..
s
Oxidation Rates and Rehabilitation Design.
[1]
where Ec is the mass of oxygen used per mass of fast-reacting, acid-consuming material (0.6
for calcium carbonate), p, is the density of fast-reacting, acid-consumed material in the wastes
(kg m-
3
), and S' is the intrinsic oxidation rate {kg (0,) m-
3
s-
1
).
Similarly the timescale to oxidise the sulfidic material is
p"'
's - --s--
[2]
where is the mass of oxygen used per mass of reactant in oxidation reaction (1.75 for sulfur),
and p, is the bulk density of reactant (kg m-
3
).
It is clear that for typical values of the lOR in waste rock (Table 1 ), and for typical values of
Pc and p,, the timescales for 'c and 1:
5
are years and hundreds of years respectively.
Table 1. Measured Intrinsic oxidation rates for different rock types at various mlnesltes.
Minesite Entity
Rum Jungle, Australia waste-rock dump
Norwich Park, Australia coal rejects
Woodlawn, Australia
Aitik, Sweden
Aitik, Sweden
waste-rock dump
waste-rock dump
large columns
Intrinsic oxidation rate Reference
(kg (0
2
) m-
3
s-')
(0.3 to 8.8) x 10_. Harries & Ritchie (1981)
(0.3 to 2.2) x 10 _. JW. Bennett & Y. Tan;
pers. comm.
(0.2 to 2. 7) x 1 o-
(0.3 to 4.3) x 1 o-
1.4 X 10-
8
Y. Tan; pers. comm.
Bennett et al. (1994)
Bennett et al. (1994)
In this context it is also useful to have some idea of water-borne pollutant transport rates through
the waste material. Putting aside the possibility of preferred paths, the relevant timescale is
1 ~ EJ- (3)
w ow
where cw is the fractional volume of the dump which is water filled (typically 0.1 ), Lis the dump
height, and Ow is the infiltration rate (m s-').
It follows that, for a typical dump heights and infiltration rates, this timescale is also of the order
of years. In practice the time for drainage, characterised by low pH and high trace metal levels,
to appear at the base of a dump is a convolution of the time to use up fast-reacting, acid-consuming
material, and the time for pollutants to be transported through the dump. From the above, the
timescale is the order of years, and will be much shorter than this only if the lOR is high, or
if the density of fast-reacting, acid-consumed material is low or if preferred paths dominate pollutant
103 RITCHIE
Oxidation Rates and Rehabilrtation Design.
transport. Pollutant levels increase with time as different parts of the dump start to contribute
to the overall pollutant load, this timescale being a reflection of heterogeneity of the dump. Pollutant
levels reach a peak, after which they decrease slowly with time, the timescales being the order
of 't
5
This is illustrated in Fig. 1.
LM ~ - - - - /- I ~
Load or
concentration
of
pollutant in
drainage
Symbol Description
tl tpK
Time
~
Time to first appearance of acid drainage
\pK
Time when pollutant concentration/load
reaches a maximum
tE
Time when drainage is no longer polluted
tE
Range
Weeks to years
Years to tens of years
Tens to hundreds of years
!_,. Maximum pollutant load/concentration e.g. 100- 100,000 mg L-' S0
4
Fig. 1. Schematic time dependence of pollution In drainage from mine wastes.
The lOR can also be used to estimate the pollutant loads and pollutant concentrations in drainage
particularly if the lOR is low. If we assume for simplicity that somehow the whole volume of
the dump is oxidising at the same rate then the overall oxidation rate, or GOR (kg m-' s-'), is
GOR- s L [4)
and the concentration (Cp) in drainage after a time, which is the order of 't.,. is
[CJ _ rp s L
ow
[5]
where rP is the mass of pollutant produced per mass of oxygen consumed (1 .71 for sulfate).
In most large waste rock dumps the larger the lOR, the smaller the region where the oxygen
concentration is high enough to support oxidation unless some advection or convection enhances
gas transport through the dump. Hence expressions [4) and [5) must be used with some caution
as they may overestimate loads and concentrations .. As a rough guide they are applicable if
the lOR is $ 1 0_. kg {0
2
) m-3 s-', and a dump $ 20 m high. This point is dealt with in more
RITCHIE 104
Oxidation Rates and Rehabili1ation Design.
detail in the Section 4 below. More detail on timescales, expected temperature rises, and expected
pollution loads are given by Ritchie (1994a, 1994b).
Table 2 presents some timescales and pollutant concentrations expected for a range of lOR
values on the assumption that the dump is aerated throughout. Some entries are omitted where
they are unrealistically large or small.
3.0 TYPICAL VALUES OF THE lOR
The lOR can be inferred from measurements of the oxygen profiles or from temperature profiles.
Table 1 presents some results from measurements at different mine sites. More recent
measurements (see for example those for Mt. Lyell discussed below) confirm that the data in
Table 1 are typical of results for a large range of waste rock dumps and for different waste rock
types.
4.0 SIMPLE MODEL
Gibson et al. (1994) described a model of oxidation in a waste rock dump which provides simple
expressions that can be used to estimate such properties of interest as pollution loads and the
time dependence of pollutant loads, using readily available data. The model assumes that diffusive
transport of oxygen dominates which is a reasonable assumption for large waste rock dumps.
It also assumes that the lOR does not change with decreasing oxygen and sulfur concentrations.
As detailed by Gibson et al. (1994), this is a reasonable assumption for waste rock dump material.
More recently, Ritchie (1994c) has shown that the assumptions can lead to an overestimate
of pollutant load, but the overestimate is only about 60%.
In view of the simplicity of the expressions, and the insight they provide to oxidation in a mass
of pyritic material, a brief description of the model is reproduced here.
It follows from the assumptions on gas transport and functional form of the lOR that oxidation
is initially confined to a region at the top of the dump. The bottom edge of this region is denoted
by x,. There is no oxidation below x,, because below x,. the oxygen concentration is zero.
This situation pertains for some long time (typically tens to hundreds of years) until all of the
sulfidic material in the region is used up. There is then a short transient condition (typically
lasting weeks) after which oxidation is confined to a region between x, and x,. In the region
0 $ x $ x,, there is no oxidation because there is no sulfide left; in the region x 2 x
2
, there is
no oxidation because there is no oxygen. This stepwise movement of the oxidation regions
continues until all of the material has been oxidised. Figure 2 exemplifies the oxidation rate
in a dump for typical dump values.
The positions xn are given by
Xn ~ fnX,
[6]
where
X ~ j 20C0-
' --
s
[7)
and where C
0
is oxygen density of air (0.265 kg m-
3
).
105
RITCHIE
Oxtdation Rates and Rehabilitation Design.
a:
Q
Q)

0
Q)
"0
E
c:
Ol
ro
E
Q)

0
Q)
0
c:
ro
0
:;:::
c:
Ol
Cii
Q)
.c:
1-
C\i
Q)
:0
Ill
1-
RITCHIE
c

c<
8_J
::JOJ
o.S
0 ;::-

0
c Ol
(/)8.S
ci
E
2
X E
0 Q) Q)
>-

$?
c
0
'7
-Cll E
<llcCll \Ol
I
Q) Cll(il$: C\1
Q)
Ul


E
i=%8.
fDQ)
::J -
0 (1j
; l5 1"0

1- ::J 0 C\1
1'--
<D
0.01(
a>::J2
E Cll"C
i= 6::Si 0
QJ
1
0
w"'
o'
c Ul
Cll'?
iii E
Q)OJ
a:2S
a:
Q
c
0

"0
x
0
ou
:;:::;Q)
o en
:o.:g
.2
c 0 c:
a>(ilQJ
a> E
Ul o.c
2 2i.

"I
0
Ol
C'.i
"0
Ol
0
:::
"<t
('")
I'-

c
.2
{ij(ij
x .2
o E
oO>
- .c:
.Co
c
- 0
u-
{g * .
:3
c :;::;
"(11
* O:.c:o
'I
0
Ol
"!
:::
cq
>,
<D
<D
0
":
<D
QJ
1
.c:
.Ql .
0

0
0

C;j
0
Ol
C\1

0
:::

(")
>,
I'-
ll)
<D
1'--
<D
QJ
u;
(1j
:::
.c: en c
0. -
2!:- E -o
0> ::J c
2-a
to=.::.::: -
g 2 8.
'- E
g;.E:!
:g
a>3; O.u
m

0 0
106

0
<D
N"
("')

0
0
>,
C\1
li)
C'.i
>,
1'--
tri
<D
0
I'-


"0
c
:J

2 E
!1:1 ::J

:;::.,
0 0
- 0

o. U)
>,!1:1
1- :;::
"'
0
("')
-.i
<D
C\1
>,
"!
li)
C\1
>,
I'-
ll)
<D
0
0
1'--
<0

- c
0> 0>
E
0> c
0
.c =
0 >
- c
.c: 0>
:;::_ .
-CIIE
ro w
Q)
(ilroo
a: EO.
0
b
ci
E
::J
"0
"5
0
.c:
OJ
::J
0

"0
Q)
c
"iii
E
"iii
E
Q)
1
c
0

"0
x
0
<j"
E
'>-
E
li)
2-
' en
E
"'
0
'1 X
EO>


OJ li) c
.!.::
0 Cil
cry O!::o
Q) -= t;::
E
>. c -
0. - E
(/) g 0 .
co :.;::;{\J"':'":

(/) 0 c 0>

-o
0
c.c:ro
'-.._..0
2'1 E
:; EON ::J "0
Ul "0 c
. - (J) . - -
'1 -"'cq-o c


ro---
ooo.c c
o-(I)OID
l{)(/) 00
c Q) c
y-W'-Uo
>.UOO>o

em wenD

..::.:::: -e (/) .., ..,
:SroroOO
roUCJCI)CI)
< co 0 0 w
Oxidation Rates and Rehabilitation Dssign.
">-
f:
"'

OJ

(!)
'
c:
0

u
")(
0
a.
E
::J
0
15
10
50 100
Time (years)
lOR
10-7
sx1o
10"8
150
Fig. 2. Overall dump oxidation rate as a function of time.
The GOR is given by
GOR - ficaos ([n -
GOR- J2C
0
os J s t]
ep,.
200
where r is time (s). The time to oxidise each region is the same, and is
E P,.
't,- --
s
The time T
0
to oxidise all the material in the dump is
T
0
- M,
where XN_, < L < x"
[8]
[9]
[1 0]
[11]
As D does appear to vary by more than a factor of about 2 or 3 in unsaturated waste dump
material (Ritchie 1 994a), the main variable in expressions [7] and [8] is the lOR. If in expression
[7] we use a typical value of D of 5 x 10-6 m
2
s-', then the oxidation would take place in all of
a 15 m high dump if the lOR = 1.1 x 1 o-a kg (0
2
) m""" s-
1
This means that the whole dump
contributes to the GOR. The oxygen concentration decreases with increasing depth within the
dump and is just zero at the bottom of the dump. For a larger value of the lOR, a smaller region
of the dump contributes.
Generally, interest focuses on the stage when the GORis greatest. This is when the oxidation
region extends in from the top surface, that is when n = 1 in expression [8]. It follows from
expression [8] that the pollution load from a dump increases only as the square root of the lOR.
107 RITCHIE
Oxidation Rates and Rehabilitation Design.
For large values of lOR, expression [8] can be further simplified to
GOR - J OCo Pos
21"
[12]
Note that now the GOR is independent of the size of the lOR, the reason being that, as the
lOR increases, the region where oxidation occurs shrinks, and the GOR is limited by the rate
that oxygen can diffuse through to this thin region. Figure 2 indicates that, except near time
zero, this type of behaviour already applies when the lOR is greater than about 1 o-' kg (0
2
)
m-
3
s'. From a practical point of view, 10-' kg (0
2
) m-
3
s' is an infinitely high lOR.
A cover of inactive material of thickness xc and diffusion coefficient D, modifies expression [8]
to
GOR- J2C
0
DS' (ra:;;; - Ja+n-1)
[13]
where
( )
2 ( )
xc s o
a- De 2Co
(14]
The case of interest is n = 1, that is when the oxidation region is just under the cover and the
GOR is greatest. The only difference between expression [14] and expression [8] is the appearance
of the term a. It is clear then, that for the cover to have any impact, a must be bigger than
unity. This is easily achieved, but means that D, has to be about 100 times smaller than D.
When a is much larger than 1, the expression [14] simplifies to
GOR- CoDe
[15]
XC
This is just the oxygen flux through the cover when the oxygen concentration below the cover
is much lower than it is in air.
5.0 A FIELD EXAMPLE
5.1 Description of Waste Rock Dump
The field example is a waste rock dump at the Mt. Lyell mine site in Tasmania. The dump (see
Fig. 3) is about 25 ha in area, abuts onto a hill along most of its eastern edge, and ranges in
depth from zero at the eastern edge to about 90 m. Table 3 provides data on the dump and
the climatic regime. The dump is some 40 years old and appears to have been constructed
by end dumping of material from an open cut immediately to east of the dump. As can be
seen from Fig. 3, the dump has a relatively flat top, has another, smaller flat some 15m below
the top, and slopes steeply to the west and north-west. It appears that the dump forms the
bulk of a catchment for a seepage channel which flows west from the western most toe of the
dump.
RITCHIE
108
i:
I:_
' I
"\.
Oxidation Rates and Rehabilitation Oes1gn
t! ..'..j
, ..
\ . ~
\\. \ .
i'j. __ ,,
!.
1:\
',,
r/'f I .
1 i
"
*
!? - ~ / .
rli' 'f.
.......
'c'
\
/
I.
\--.
/,:
/. j
- .,
!
Fig. 3. Map of the Mt Lyell shaft waste rock dump, showing ANSTO monitoring liners.
5.2 Probe Holes
Nine holes were drilled into the dump and probe tubes (Bennett et aL 1994) inserted to allow
measurement of temperature and oxygen profiles through the depth of the dump. Where possible
these holes were drilled through to the material underlying the waste. The position of these
109 RITCHIE
Oxidation Rates and Rehabilitation Design
Table 3. Data applying to Mt Lyell shaft
waste rock dump.
Parameter Value
Area of dump
Mean height of dump
Mass of dump (estimated
volume; total tonnage
extracted from West Lyell
open cut was 47 Mt.)
25 ha
50 m
20 Mt
Sulfur content 5%
Dumping started (effectively) 1945
Dumping ceased 1952
Mean annual precipitation at 2,500 mm y'
Mt Lyell
Average ambient 13 oc
temperature at Mt Lyell
Range in ambient
temperature at Mt Lyell
2-23 oc
Table 4. Dump depth and depth lined at
the probe hole positions.
Hole
WRC001
WRC002
WRC003
WRC004
WRC005
WRC006
WRC007
WRC008
WRC009
Dump depth
(m)
75.0
92.0
47.5
30.0
36.0
48.5
34.0
68.0
42.0
Lined depth
(m)
32.4
90.0
47.5
5.0
17.5
48.5
34.0
23.0
42.0
probe holes, WRC001 to WRC009, are
shown in Fig. 3, and some details of the
dump depth and the depths of monitored
waste given in Table 4.
5.3 Results and Interpretation
The temperature profile in WRC005
(Fig. 4a) shows a seasonal change in
temperature near the top of the dump
which decreases in amplitude with depth
until it is vanishingly small at about 6 m
from the surface. This is a typical
temperature profile when there is little
or no oxidation, and consequent heat
generation within the dump. If there
were no heat generation, then the
temperature below about 6 m would be
close to the annual mean which is 13 oc
at Mt. Lyell.
The relatively constant oxygen concen-
tration profile is consistent with little or
no oxidation in most of the dump in the
region of WRC005. The small decrease
in the oxygen concentration between
2 m and 6 m is consistent with some
oxidation in the region, as is the slight
elevation of temperature above the
mean of 13 oc in much of the dump
below 6 m. Both temperature and
oxygen profiles are consistent with a
very low oxidation rate in material in the
part of the dump near probe hole
WRC005. The elevated temperatures
at WRC001 (Fig. 4b) and the oxygen
profiles are consistent with a significant
oxidation rate throughout most of the
dump in this region, and with an oxygen
transport process dominated by diffusion
in the top 10 m or so, but by some
advective process in the lower part of
the dump. This picture is very similar
to that found in parts of the waste rock
dumps at the Rum Jungle mine site
(Harries and Ritchie 1981) where the advective process was identified as convection.
Similar mechanisms are at play in the region of WRC003 (Fig. 4c). There are two further points
of note here. The considerably elevated temperatures seen from 10 m to 30 m together with
high oxygen concentrations indicate both a significant oxidation rate and an effective gas transport
mechanism in this part of the dump. This gas transport mechanism has not yet been identified,
but it could be the result of wind blowing over the dump. Winds at Mt. Lyell are consistent in
RITCHIE
110
a:WAC005

,,
g
8 ..

g:,
0


0

., .
f-
.
Oepth(m)
c:WAC003
p(

l
E.
"'
. --1
10 iO )Q
Deplh{m)
c
g


810

0
l!O
c
.2

&.o


Oxidation Rates and Rehabilitation Design.
b:WAC001
.
Depth (m)
d:WRC006
I--......J4---+- ,-- - I
Depth (m)
Fig. 4. Oxygen and temperature profiles for locations WRCOOS (a), WRC001 (b), WRC003
(c), and WRC006 (d).
111 RITCHIE
Oxidation Rates and Rehabilitation Design
both direction and strength, both of which are required for wind-induced transport. The second
point is the very rapid decrease in oxygen concentration from about 4 m to 6 m, followed by
a rapid rise from about 10m to 11 m. The profile and the time evolution of the profile in this
5 m to 6 m large region are both consistent with a pod of material of high oxidation rate, either
at this point in the dump or very close to it. The set of profiles at WRC006 (Fig. 4d) is another
example of a similar situation. In this case, the oxidation rate of the material below the pod
is not as high as in the case of WRC003.
5.4 The Range of Oxidation Rates
Analyses of the time evolution of the oxygen profiles after installation of the probe holes and
of the temperature and oxygen profiles characteristic of pseudo steady state oxidation in such
a waste rock dump, provides estimates of the oxidation rate at various points in the dump. The
picture which emerged was one where there were pods of high oxidation rate material embedded
in a large mass of material with a much lower oxidation rate. The vertical extent of these pods
was typically a few metres. The oxidation rate associated with the pods was about 6 x 1 o'
kg (0
2
) m-3 s', and that with most of the rest of the material about 6 x 10-
9
kg(0
2
) m
3
s'.
6.0 POLLUTANT GENERATION RATES
Expression [7] indicates that gas transport by diffusion can support oxidation over a region of
some 21 min depth when the diffusion coefficient has the typical value of 5 x 1 o...s m
2
s' (Ritchie
1994a), and where the oxidation rate is 6 x 1 o-
9
kg (0
2
) m
3
s'. It must be noted that if diffusion
is the only gas transport mechanism, the oxygen concentration will decrease continually over
this 21 m region. A feature of the oxygen profiles at the Mt. Lyell waste rock dump is the near
atmospheric level of oxygen except were there are pods of high oxidation rate material. It is
easy to show, using a code such as FIDHELM (Pantelis and Ritchie 1993; Pantelis 1993) that
a pressure difference of only a few tens of Pascals across a 50 m high dump can maintain a
high enough gas flow to support an oxidation rate of 1o kg (0
2
} m' s' at a gas permeability
typically found in waste rock dumps (Ritchie 1994a). Such pressure differences can readily
be generated by wind flow over a dump of the shape of that at Mt. Lyell. It is therefore reasonable
to accept that the high oxygen concentrations are supported by a mixture of diffusion and advection.
It follows that the overall oxidation rate (kg s') in most of this dump is just the product of the
inferred lOR and the volume of the dump in which this lOR pertains. The overall pollutant
production rate from such regions of the dump can then be readily estimated.
The results in Section 4 show that diffusive transport can support oxidation over a distance of
only about 2m when the oxidation rate is 6 x 10' kg (0
2
) m' s'. The pressure differences
required to support oxidation at these high levels by advective gas transport are too high to
be achieved by natural means such as wind effects. It follows that the contribution to the overall
oxidation rate from a pod of high oxidation rate material comes only from a comparatively small
region. If gas transport mechanisms maintain comparatively high oxygen concentrations around
the pod, then the contribution will be from a thin skin (-2m thick) around the pod.
The total length of the probe holes was 295 m, of which 23 m can be taken to be reacting at
the high rate. The results from the probe holes applied to the whole dump implies that 8% of
the dump volume is contributing at the high rate. Using data on the total annual sulfate load
from the dump, and assuming that this comes predominantly from the high reactivity regions,
leads to the conclusion that some 13% of the dump volume is contributing to the load. Given
that the probe holes sample a comparatively small part of the dump, this agreement is quite
good.
RITCHIE 112
Oxidation Rates and Rehabilitation Design.
7.0
MANAGEMENT STRATEGIES
7.1 Estimates of Current Status
In broad terms, there are three options for managing pollution from an existing waste rock dump:
(1) Do nothing except manage the polluted drainage in some way;
(2) Speed up the oxidation process to exhaust sulfides within a reasonable timeframe
and manage the polluted drainage in some way; and
(3) Slow down the oxidation rate to an extent where polluted drainage requires minimal
or no management.
To assess the practicality and efficacy of these options, it is useful, often crucial, to have a
picture of where and at what rate pollutants are generated within the waste rock dump.
A picture of the waste rock dump at Mt. Lyell that is consistent with the measured properties
is that:
(1) There are pods of material embedded in the dump; the oxidation rate in the pods
is about 6 x 10
7
kg (0
2
) m-3 s';
(2) The rest of the material in the dump has a much lower oxidation rate of 6 x 10"
9
kg (0
2
) m
3
s' at most;
(3) Although these pods vary in size, a reasonable picture is one of some 600 pods
each of 1 0 m diameter;
(4) It is an outer 2 m thick rim of these pods that contributes the bulk (about 90%)
of the pollutant load generated by the dump;
(5) Gas transport mechanisms in the dump are such that generally high oxygen
concentrations prevail everywhere except within the pods;
(6) The time to oxidise the sulfidic material in a pod is about 20 years; and
(7) The time to oxidise the sulfidic material in the rest of the dump is about 600 years.
On the basis of this picture the 'do nothing' option implies the current pollution load and pollutant
concentrations continuing for a further 20 years. At that time the load and concentration start
to drop to a level of about 10% of the current level and continue at that level for a further 600
years. The estimate of 20 years is a minimum since it assumes that the sulfur content of the
pods is the same as the average for the whole dump. A decrease in pollutant concentration
in drainage by a factor 10 from current levels implies sulfate concentrations in the range of 1,500
mg L-' to 3,000 mg L-'. and copper concentrations in the range 10 mg L' to 30 mg L'.
The 'do nothing' option requires careful management of drainage both in the short term (20 years
or more) and in the long term (about 600 years).
113 RITCHIE
Oxidation Rates and Rehabilitation Design.
7.2 Speed Up the Oxidation Rate
Currently the rate of overall oxidation in the pods is limited by oxygen transport rates into the
pods from a high oxygen concentration regions surrounding the pods. It is clearly possible in
principle to locate these pods and supply air to them in such a way as to exhaust the fast reacting
sulfidic material quickly. Speeding up the oxidation rate to the rest of the dump is much more
of a problem because: (a) there appears to be little or no rate limitation due to oxygen supply;
(b) the material has been there for many years which is more than adequate for the bacterial
ecosystem to adapt to conditions and multiply; and (c) the measured rate is consistent with that
measured for the bulk rate in many dumps (see Section 3).
It would appear then that measures to speed up the oxidation rate would decrease the timescale
for oxidation of the pods but do little or nothing to change the timescale of 600 years or so
associated with pollutant concentrations in drainage which are of a level to require management.
7.3 Slow Down the Oxidation Rate
The simple analysis presented in Section 4 indicates that a cover with a low gas diHusion coefficient
can reduce the overall oxidation rate in a dump by a substantial factor. The factor depends
very much on the properties of the cover, and hardly at all on the properties of the dump material,
including the magnitude of the lOR. From this point of view, this option is attractive, the practical
problems are the design of a suitable cover, and the integrity of the cover over a very long
timescale.
8.0 ASSESSMENT OF PERFORMANCE
Since the primary pollutant production process in pyritic waste is the oxidation of sulfidic material,
a measurement of the overall oxidation rate in the dump is a measurement of the effectiveness
of schemes to reduce pollutant production in the wastes. It is not the intention in this paper
to describe the details of how the overall oxidation rate can be measured, rather it is the intention
to emphasise some basic principles, particularly those which apply when a cover system is used.
In a dump where a cover system is installed to reduce the flux of oxygen into the dump, the
oxygen concentration below the cover is an indicator of cover effectiveness. It should however
be stressed that such a measurement is a rapid way to monitor when and if the cover is Ineffective
since anything other than very low oxygen concentrations (less than or about 0.1% mole fraction)
indicates that the cover is not working, or for some reason the oxidation rate of sulfidic material
is very low.
Generally the oxygen concentration below the cover is so low that compared to the 22%
concentration in air it can be taken to be zero. The flux through the cover is then simply
F ~ DCO
[16]
XC
and the magnitude depends primarily on the magnitude of the diffusion coefficient. In practice
an estimate of the overall oxidation rate then devolves down to measuring the diHusion coefficient
of the cover material.
RITCHIE
114
Oxidation Rates and Rehabilrtation Design.
It is possible to carry out a direct measure of the flux through the upper surface of the pyritic
material, and hence obtain a direct measure of the overall oxidation rate. Such techniques
are still in the development phase but would seem to provide a method for assessing the flux
over a large area in a relatively short period of time.
9.0 ACKNOWLEDGMENTS
An understanding of the interactive processes which determine the pollutant generation rate
in waste rock dumps and similar structures requires input from both field measurements and
quantitative mathematical modelling. John Bennett and Andrew Garvie backed up by Alan Boyd,
Norm Clark, Arthur Dixon, Warren Hart and Viphakone Sisoutham have developed novel field
measurement techniques and produced high quality measurements. David Gibson, Garry Pantelis
and William PlotnikoH have developed calculational tools which provide insight into the interactive
mechanisms and the relative importance of data characterising dump material. We have had
considerable assistance from personnel at a large number of mine sites in installing equipment
and carrying out routine measurements. Finally I wish to thank both the Mt. Lyell Mining and
Railway Company and the Tasmanian Department of Environment and Land Management for
use of data from the waste rock dump at Mt. Lyell.
10.0 REFERENCES
Bennett, J.W., Gibson, D.K, Ritchie, A.I.M., Tan, Y., Broman, P.G., and Jonsson, H. (1994).
Oxidation rates and pollution loads in drainage: correlation of measurements in a pyritic
waste rock dump. Proceedings of the International Land Reclamation and Mine Drainage
Conference and Third Conference on the Abatement of Acidic Mine Drainage, Pittsburgh,
USA, 1994. 1, 400-409. (USA Department of the Interior, Bureau of Special Publications,
SP 06A-94)
Bennett, J.W. (1989). In situ air permeability measurements in porous material. 81 st Annual
meeting of the Soil Science Society of America, Las Vegas, Nevada, USA, October
1989.
Blackford, M.G., and Harries, J.R. (1985). A heat source probe for measuring thermal conductivity
in waste rock dumps. Australian Atomic Energy Commission Report, AAEC/E609.
Davies, G.B., and Ritchie, A.I.M. (1986). A model of oxidation in pyritic mine wastes- I Equations
and approximate solutions. Applied Mathematical Modelling 10, 314-322.
Davies, G.B., and Ritchie, A.I.M. (1987). A model of oxidation in pyritic mine wastes- Ill import
of particle size distribution. Applied Mathematical Modelling 11, 417-422.
Gibson. O.K., Pantelis, G., and Ritchie, A.I.M. (1994). The relevance if the intrinsic oxidation
rate to the evolution of polluted drainage from a pyritic waste rock dump. Proceedings
of the International Land Reclamation and Mine Drainage Conference and Third Conference
on the Abatement of Acidic Mine Drainage, Pittsburgh, USA, 1994. 2, 258-264. (USA
Department of the Interior, .Bureau of Special Publications, SP OGA-94).
Harries, J.R., and Ritchie, A.I.M. (1981 ). The use of temperature profiles to estimate the pyritic
oxidation rate in a waste rock dump from an opencut mine. Water, Air and Soil Pollution
15, 405-423.
Harries, J.R., and Ritchie, A.I.M. (1985). Pore gas composition in waste rock dumps undergoing
pyritic oxidation. Soil Science 140, 143-152.
MillerS. (1992). Acid drainage from mine waste and mining operations- Theory and prediction
techniques. Acid Mine Drainage Workshop, Strahan, Tasmania, 1992, pp 39-89.
(Tasmanian Chamber of Mines: Hobart, Tasmania).
Pantelis G., and Ritchie A.I.M. (1990). Macroscopic transport mechanisms at rate limiting factors
in dump leaching of pyritic ores. Applied Mathematical Modelling 15, 136-143.
Pantelis G., and Ritchie A.I.M. (1992). Rate limiting factors in dump leaching of pyritic ores.
115
RITCHIE
Uxdat1on Hales and fiehabil1tation Des.lgn.
Applied Mathematical Modelling 16, 553-560.
Pantelis G., and Ritchie A.I.M. (1993). Optimising oxidation rates in heaps of pyritic material.
In 'Biohydrometallurgical Technologies, Proceedings of the International Biohydrometallurgy
Symposium, Jackson Hole, Wyoming, USA, 1993.' (Eds A. E. Torma, J.E. Weyand V.I.
Lackshmanan.) 1:731-738. (TMS, USA)
Pantel is G. (1993). FIDHELM: Description of Model and Users Guide. Australian Nuclear Science
and Technology Organisation report. ANSTO/M123.
Ritchie, A.I.M. (1977). Heap leaching: a gas diffusion rate limited model. Australian Atomic
Energy Commission Report, AAEC/E429.
Ritchie A.l. M. (1994a). The waste rock dump environment. In 'The Environmental Geochemistry
of Sulfide in Mine-Waste. MAC Short Course Handbook.' (Eds D.W. Blowes and J.L.
Jambor.) 22:133-161. (Mineralogical Association of Canada: Nepean, Ontario, Canada).
Ritchie A.I.M. (1994b). Sulfide oxidation mechanisms: controls and rates of oxygen transport.
In 'The Environmental Geochemistry of Sulfide in Mine-Waste. MAC Short Course
Handbook.' (Eds D.W. Blowes and J.L. Jambor.) 22:201-244. (Mineralogical Association
of Canada: Nepean, Ontario, Canada).
Ritchie A.I.M. (1994c). Biooxidation heaps and AMD from waste rock dumps- The importance
of the intrinsic oxidation rate. Australian mining looks north, challenges and choices.
1994 Annual Conference of the Australian Institute of Mining and Metallurgy, Darwin,
1994. pp 473-478. (AUSIMM: Melbourne, Australia)
RITCHIE
116
Second Australian Acid Mine Drainage Worl<shop (Eds. N.J Grundon and L C. Bell), pp_ 117-120
GEOCHEMICAL INDICATORS OF SULFIDE OXIDATION
AND ACID GENERATION IN THE FIELD
S. Miller
Environmental Geochemistry International Ply. Ltd.,
20 Rowntree Street, Balmain, NSW 2041.
ABSTRACT
Preliminary prediction tests and operational monitoring of mine waste materials does not guarantee
that all potentially acid generating materials will be identified. Therefore, it is advisable to maintain
a monitoring and surveillance system that will provide an early warning of any sulfide oxidation
and acid generation problems on site. This paper discusses simple field indicators of acid
generating materials and simple measurements that can be made to indicate the presence of
both acid generating and acid consuming materials. Geochemical processes which occur within
mine waste are outlined, and indicator parameters that should be monitored in rock and drainage
water are discussed.
1.0 INTRODUCTION
Acid rock drainage (ARD) prediction investigations are aimed at classifying the acid generating
characteristics of individual samples of mine rock and process wastes. Based on the results
of these investigations, geochemical rock types and waste types are identified and waste quantities
and mining schedules defined. The most appropriate operational management plan can then
be developed and implemented. This plan normally involves one or more of the following strategies:
(1) Immediate burial and isolation of potentially acid forming (PAF) material;
(2) Delayed burial and isolation of PAF material;
(3) Segregation and controlled placement of PAF material;
(4) Blending PAF with non-acid forming (NAG) or acid consuming (AC) rock;
(5) Treatment with alkali and/or bactericides;
(6) Placement of an engineered cover for infiltration control, oxygen diffusion control
and revegetation; and
(7) Placement under permanent water cover.
Even though pre-mining prediction work has been carried out and 'best practical' strategies are
implemented, it is advisable to maintain a monitoring and surveillance system that will provide
an early warning of any sulfide oxidation and acid generation problems on site. In addition,
sites with only a low risk of acid generation are encouraged to implement a monitoring system
to confirm that sulfide oxidation and acid generation are not a concern. This paper discusses
the geochemical process which occur within mine waste materials and identifies indicator parameters
which should be monitored.
2.0 GEOCHEMICAL REACTIONS AND REACTION PRODUCTS
Figure 1 presents a schematic of the primary acid generation reactions, acid buffering and
equilibrium reactions and reaction products (soluble and precipitates). The approximate pH
scale is shown and indicates the pH where the solution indicators and precipitates may first
be observed. The pH range for the two major sulfide oxidation steps are also shown.
117
MILLER
"T1
tO
G)
(()
0
0
::r
<1l
3
0
e1.
.....
(!)
n>
a
0
:I
(/)
Ol
:I
a.
:I
a.
g
0
u;
0
....
(/)
:
0:
(l)
0
><
0:
!:l.
0
:I
Ol
;:I
a.
Ol
0
0:
(Q
(l)


0
::J
MILLER
:J ;:, ::J ;:, :J
0 0 0 0 0
rororororo

0.
a_-.::J a.::J
::; s.



o:::>""Q.p;o
_ : swo
8
n>:Tcn<O"'"'

2. 0 ::T'<:: 0.:
-xOl-.:J<D
3 -(/) :J () (/)
3
- g.
0
a: c w

ii).O 0-< a.::r
- , 0 s rr n>
o m en
:l.iiJ s:: (il 0

::gnl'::


0.!0 0-o (JI z
o-uco::!.-
2. I Ul
::rrrOroron>
pj c ... ...........
(O::t:OO::)C..


;;!_ g_ 8
::rorrosro


g a-g :J
C:::l.-,::roro

g ::;a. ::la.= P3
0 :;: 0 :J :-


- "'::>x-

wO:&Ui<ll
m m ro s(l)
< '"0

g-
<T.l c5 ::J 0 0
..., (t) ro :::1
ec Ill 3 .,c
..0
'-< PJ
. CD
CD Ol
:::r 0
Ol -
(J) a.
0" ()
<t> 0
(!) :J
::r (/)
c
0 3
6: :;
o; <0
<1l
p.
c....
a
(/)

.._,
0?
f>l:
(/l '"Tl'Tl

(/) Vl -, II
N
N C1> C1>
<flU>
N
+
-n +
-"'
0
+ N
rS'!
+
+ +
+
j
I

N
0

--

-n
-(I)
+
-
- - -
"'
+ ..
0 "'

"Tl
+ C1>
:r:
"'+ <8.
;r 0
+ +
p+
:t c; 23
+
0
... l!!.
w
"Tl 0
0
+ +
il;" ....
"'+ 1!:. C1>
Pf
:t
+ fr
:t g
+ +
g
l;-!J

-4- - -
-
N w ... (/1 Ol
.._,
(X)
:r:
I
7<'- +
...,.,

+ C1> ....




3
mg o
t5>88::,
t.

- .,
"' "'
'<
t


"' ::.



"Tl
Q
:j: :t
" Jl!-!.1
+
0
cr.
=
0
23 <l>N>'<

0 0
w

oN
w (Q
N w ...
(/l
""
...,.
"'
Increasing concentration (I)
0
-4- - - - - - - - -:-
-n
-n
n
"'
C1>
"Tl -::;;

C1>
C1>
+
+ N
+
0
+
+ )> c:
"'
?'
..... + -
"Tl
..... :;:::
C1>
.,
-"' +
C1>
l>:J
"'0
+
+
- ..
2:
+ ?.q
<0
0 0.<1>
' lll <
.s

u 3

Io
"'
w
N w ... (/l m
.._,
0>
o-n

; Q
"Q 0
)><... n>23
_.,
8
(/}
O.w
c: .... 0

" 0
...
...
;;: !!!.
:..., N
N
C1> ...
=
:r:
C1>
N N
0 0
i

..

118
<D

-
"
;o

;o
-<
;o


0
z
V>
"'
"&
""
!:ii
"Tl
m
;o
z
C)
;o
m
)>

0
z
Ul
"'

(/)
0
r

0
z
z
0
n

0
"'
(I)
"'

-o
;o
m
!J
-o
=1
)>
-;
m
Ul
Geochemical Indicators ot Sulfide Oxidation.
These are:
(1) Direct air and water oxidation; and
(2) Indirect oxidation by ferric iron {catalysed by bacteria).
As indicated on Fig. 1, iron sulfide minerals such as arsenopyrite, chalcopyrite, pyrrhotite and
pyrite are the main acid-generating sulfides with pyrite the most common contributor. Non-ferrous
metal sulfides such as covellite (CuS). galena {PbS), millerite (NiS) and sphalerite (ZnS) are
not generally acid generating and sulfides such as chalcociie (Cu,S) are acid consuming.
Early indicators of sulfide oxidation in mine site drainage are increasing SO, (and frequently
an associated decrease in the Cl to SO, ratio), increasing concentrations of major cations and
a shift in pH. In some cases the pH can increase significantly due to the acid neutralising reactions,
and pH values as high as 9 can be recorded. Low pH values can also be recorded early if the
buffering capacity of the site is low.
Other important early warnings can be provided by Mn and Zn. Increasing Mn concentrations
are observed even at neutral to slightly alkaline pH where rhodochrosite occurs in association
with iron sulfides. Zinc is also frequently the first metal observed in leachates. Copper follows
soon after and can be recorded at high concentrations even at pH values of 6.
Aluminium is normally observed at pH values below about 4. Dissolved iron can be observed
over a wide range in pH. Ferrous iron can occur in seepage at high pH if low redox conditions
exist, but the iron will rapidly oxidise and precipitate from solution where the seepage daylights.
At low pH both ferrous and ferric remain in solution, and very high concentrations can be observed
at pH values below about 3.
The ferric to ferrous ratio tends to be low at pH values greater than about 3 and below 3 the
ratio can be high.
Ferric oxide and hydroxide precipitates can be observed over a wide pH range. Co-precipitates
such as arsenic ferrihydrite can also be observed at high pH, and this is an important process
for the removal of dissolved arsenic from mine site drainage. Precipitation of gypsum is commonly
observed in mine waste where sulfides and calcites occur and relatively dry conditions exist
tor at least part of the time. Gypsum precipitation generally occurs when the solution concentration
of SO, exceeds about 2,500 mg L _, and dissolved Ca approaches about 400 mg L-'. Gypsum
tends to limit the upper concentration of dissolved SO, at higher pH values whereas, at low
pH, the SO, concentration can be very high exceeding 10,000 mg L _, .
Epsomite (MgS0,.7H,O). gypsum and alunite {KAI
3
(SO,),(OH)
6
) are common white precipitates
which form as a result of the acid generation and acid neutralisation reactions. Epsomite is
normally only observed where strong evaporative conditions exist such as on rock faces and
dump surfaces. Both gypsum and epsomite can occur over a wide pH range whereas alunite
tends to occur only at pH values less than about 4.
Jarosite (KFe3 (S0,)2{0H)6) is a diagnostic yellow precipitate and occurs at pH values of about
3.5 and lower. Jarosite is essentially a temporary storage for acidity and when re-dissolved
releases this stored acidity according to the following equilibrium reaction.
KFe3(S04 ) 2(0H)
6
+ 3H
2
0 H K' + 3Fe(OH)
3
+ 2SOt + 3H'
[1]
119
MILLER
Geochemical Indicators of Sulphide Oxidation.
The metastable jarosite and ferric oxides/hydroxides tend to coat soil particles and precipitate
within the voids. These minerals set up a strong pH buffering mechanism that can dominate
the solid-solution chemistry and maintain the pH at about 3 even if all pyrite is oxidised or pyrite
oxidation has been controlled. Test work has shown that exclusion of oxygen does not interfere
with reaction [1], and there.fore buried or covered acid mine waste which contains significant
jarosite will continue to release acid by jarosite dissolution for a long period of time. It is therefore
important to control acid generation in PAF waste materials that will subsequently be buried
to enable effective control of acid drainage and minimise potential post-mining liabilities.
3.0 RECOMMENDED MONITORING STRATEGY
As indicated on Fig. 1, there is a lot more to identifying sources and impacts from ARD than
measuring pH. However, the pH does provide an indication of the parameters that should be
measured and monitored over time.
pH and S0
4
are key parameters when the pH is slightly alkaline to slightly acidic. Depending
on the specific site mineralogy, As and Mn should be included along with the major anions and
cations even when the pH is not acidic. The SO/CI ratio is a good indicator of an input of sulfate
water and should also be included.
There is little value in including Fe or AI in the monitoring suite until the pH drops below about
4 unless low redox conditions prevail and ferrous iron occurs. However, Zn and Cu should be
included at pH values below about 6 to 6.5 if Zn and Cu minerals occur within the waste.
Other metals can be included, as required, depending on whether or not the metals are significantly
elevated within tha rock. To provide a measure of the extent of element enrichment multi-element
assay results for rock samples can be compared to the average crustal abundance for that element
by calculating the Geochemical Abundance Index (GAl) which relates the actual concentration
with the crustal abundance on a log 2 scale. The GAl is expressed in 7 integer increments
(i.e. 0 through to 6, respectively), where a GAl of 0 indicates the element is present at a
concentration similar to, or less than, average crustal abundance; and a GAl of 6 indicates
approximately a 1 00-fold, or greater, enrichment above average crustal abundance. The overall
purpose of the GAl is to identify any elements (especially metals) that occur at concentrations
which are well above normal background values and which therefore warrant further examination
to assess their environmental significance. As a general rule, a GAl of 3 or greater signifies
enrichment to a level that warrants further examination.
4.0 CONCLUSION
It is important to have an understanding of the physical and chemical process involved in sulfide
oxidation and acid generation to develop an environmentally secure waste management plan
for the control of acid generation and acid drainage.
This paper attempts to provide those involved in mine waste management and mine site
environmental management with basic geochemical information for assessing or setting up water
quality monitoring programs for ARD. The water quality data must be regularly evaluated to
identify trends and should be supported by geochemical acid-base and NAG test data on waste
types within particular catchments. Flow data should also be collected so that sulfate loads
can be calculated. Mineralogical and multi-element data is required to identify potential
contaminants and, if necessary, the oxidation status of existing dumps can be assessed by
measurement of oxygen and temperature profiles.
MILLER
120
Second Australian AcJd Mine Drainage Workshop (Eds. N.J. Gwndon and L.C. Ball), pp. 121-126
WATER FLOWS IN MINE WASTE DUMPS AND
THEIR IMPLICATIONS FOR AMD CONTROL
L.M. Gilbert
AGC Woodward-Clyde Pty Ltd, 49 Park Road, Milton, Old 4064.
ABSTRACT
Water flows into, out of, and around pyritic mine waste dumps and tailings dams are important
in both the generation of acidic leachate and its transport into the downstream environment.
This paper examines water flows in waste dumps and tailings dams in relation to various
hydrologicaVhydrogeological conditions which typically occur on Australian mine sites. Conceptual
models of the internal movement of water through waste dumps and tailings piles are outlined
and the implications for design are discussed.
It is postulated that internal water flow in a well constructed mine waste dump is governed by
preferential flow paths. The hypothesis has direct implications for the longer term migration
of oxidation products from waste dumps.
1.0 INTRODUCTION
Water by virtue of is movement through and around mine waste dumps is the transport mechanism
for contaminants generated in waste dumps to the downstream receiving environment. An
understanding of the hydrology of waste dumps, and in particular how water movement occurs,
is therefore necessary for a rational approach to the control of acid mine drainage.
The underlying message behind this paper is that control over the rate of migration of contaminants
from waste dumps is one of the main engineering and management options available for the
practical management of acid mine drainage. The control over rate of migration can be achieved
by controlling water flows.
The structure of the paper is to present a general overview of typical surface and groundwater
regimes common to a range of waste dump and tailings dam situations. The range of engineering
controls which can be employed to control water II uxes through dumps are then discussed and
finally a preferential flow path model, which is applicable to infiltration of water through mine
waste dumps, is described. The implications for the design of mine waste dump rehabilitation
and the control of impacts arising from oxidation of pyritic wastes are then outlined.
2.0 HYDROLOGY OF MINE WASTE DUMPS
Mine waste dumps typically comprise a poorly graded mixture of coarse rock fragments sized
from boulders down to cobbles intermixed with a small proportion of fines. Infiltration capacity
is typically high. Rain falling directly on the dump surtace is partitioned into interception, infiltration
and runoff depending on the texture, grading and topography of the dump surtace. Infiltrating
water migrates downward through the dump with a proportion appearing as leachate at the toe
of the dump.
Waste dumps are constructed in a wide range of climatic, physical and hydrological regimes.
121 GILBERT
Water Flow in M1ne Waste Dumps.
This variability is reflected in the hydrology and water balance of mine waste dumps. The following
examples are provided to illustrated the main hydrological features of mine waste dumps over
a range of conditions:
Case 1: Mine Waste Dumps in Flat Arid Areas
The arid to semi-arid areas of Australia typically have low relief and are characterised by poorly
developed surface drainage networks. Whilst average rainfalls are low, rainfall records are typified
by long periods of little or no rainfall interspersed with occasional periods of heavy rain. The
rainfall intensities of these events are often high and typical of rainfall intensities experienced
in wetter areas of Australia. The infrequent rainfall occurrences often result in significant surface
ponding and subsequent slow subsidence by evaporation and infiltration. Under these conditions,
infrequent but potentially environmentally significant flushes of leachate can be generated from
mine waste dumps.
Groundwater resources are often limited in both quality and quantity and often recharged from
areas distant from the site. The migration of water through porous waste dumps and its subsequent
rapid mounding at the base of the dump can act to isolate these waters from the usually high
rates of evaporation leading to localised groundwater recharge and expression of water as dispense
baseflow from the toe of the dump.
Case2: Mine Waste Dumps in Humid Regions
Most of the coastal regions of eastern Australia fall into this category. In comparison to the
arid areas, rainfall is higher and potential evaporation rates are lower. Rainfall occurs more
frequently and over a wide range of intensities. Runoff is likewise significantly higher and
concentrated within well developed drainage networks. Mine waste dumps are typically constructed
in valleys using relief to reduce cost. The typical situation is to construct dumps in small ephemeral
valleys proximal to the mining operations.
Groundwater resources are normally recharged locally and are typically of high quality, near
the surface, and more often used by others. The movement of surface water is dominated by
!low concentration in the valley floors. Runoff is rapid particularly in headwater valleys and diversion
of runoff around dumps may be a major issue. Groundwaters are often recharged from structural
features particular where these intersect creek lines. Creeks may be either influent or effluent
with groundwater flows being driven by local topography.
Mine waste dumps in these situations change surface water flow patterns and, in cases where
the dump is situated over groundwater recharge areas, there is likely to be local changes to
recharge patterns also. For relatively permeable dumps with low runoff potential, the effects
would be a general increase in recharge. The relatively high infiltration rates associated with
these dumps results in delayed drainage and increased ponding over and contact with the recharge
areas.
The movement of surface water downward through the dump would occur rapidly via 'macro-pores'.
A temporary water table would form at the floor of the dump where the original ground surface
forms a relatively low permeability boundary resulting in mounding and movement of waters
through the overlying relatively more permeable wastes at the contact. This results in a delayed
yield effect with water often appearing as surface base flow from points corresponding to the
original drainage lines at the downslope toe of the dump. The turbulent water flow which develops
in macro-pores following heavy prolonged rainfall can cause flushing of fines through the dump
and therefore maintenance or enhancement of these flow pathways.
GILBERT 122
Water Flow in Mine Waste Dumps.
3.0 ENGINEERING CONTROLS FOR MINE WASTE DUMPS
The potential for acid-generating mine waste dumps to cause environmental impacts in arid
areas is lower than in humid areas principally because of the lower average rate of water movement
As with any mine waste dump, the design and implementation of control measures needs to
be directed toward controlling the rate of migration of contaminants into the environment consistent
with the assimilative capacity of the environment and accepted environmental objectives.
The range and application of available engineering controls has been adequately described
in other papers presented during this workshop. The basic principle is to minimise the movement
of water through wastes which generate or contain containments. This is achieved by:
(1) placing these wastes away from areas where water flows are or are likely to become
concentrated (e.g. site selection and segregated placement);
(2)
{3)
by diverting surface and shallow groundwater flows away from these wastes (eog. surface
water diversions and groundwater cut-offs); and
by limiting movement of incident rainfall and runoff through wastes by covering them
with low hydraulic conductivity material.
The other strategies which can be employed is to enhance the assimilative capacity of the
downstream environment. This approach is in practical terms, however, less attractive in most
cases.
Superimposed on this are the 'reactionary' engineering controls which can be used such as
interception, treatment, controlled release. These measures do not involve controlling water
flows through or around waste dumps or tailings piles and are therefore not described.
4.0 HYDROLOGY OF TAILINGS DISPOSAL STORAGES
Tailings typically comprise well graded, fine-grained particles consisting of sand to fine silt and
clay-sized rock fragments. Tailings are deposited hydraulically and are therefore saturated upon
deposition. After deposition the tailings slurry rapidly settles to form a sloping beach.
Water liberated during this process appears at the downslope end of the beach" The deposited
tailings mass is typified by large pore water storage and moderate to low permeability with often
significant anisotropy.
Tailings are deposited in layers with progressive deposition over older layers resulting in
consolidation and drainage of pore water, which moves both toward the free ponded surface
water and downward toward the storage floor. The rate of vertical migration of water through
the tailings and into the underlying groundwater is determined principally by the vertical permeability
of the tailings and the permeability of the underlying foundation materials. The effect of seepage
from a tailings dam on the local groundwater flow regime will depend on the relative quantity
of water flow from the tailings to the total flow in the underlying groundwater system.
Runoff from tailings surfaces areas are typically high due to the low infiltration capacity of the
fine-grained, saturated tailings. In arid areas, evaporation from both free ponded water and
from the tailings beach can be high and enhanced over rates observed on lakes due to low
albedo. The water balance of active tailings storages is usually dominated by the tailings water
itself.
123 GILBERT
il
~ ~
Water Flow in Mine Waste Dumps.
To illustrate the main hydrological features of tailings disposal storages constructed in various
hydrological settings in the following two examples are given which represent opposite cases:
Case 1: Tailings Storages in Flat Arid Areas
The lack of topography in these areas typically requires construction of turkey's nest embankment
system to contain the tailings. Rainfall and runoff effects are therefore limited to incident rainfall
and water balance of the storage highly dependent on the tailings water. The migration of tailings
water occurs as horizontal flow and ponding in the surface water pond and vertically downward
movement through the storage floor. Initially later flow results in a wetting front moving downward
through the dry or partially saturated sub-surface materials. Seepage waters ultimately reach
the saturated zone and begin to form a groundwater mound and development of local hydraulic
gradients and associated groundwater movements away from the tailings storage.
The mounding and associated groundwater gradients increase until the rate of migration of seepage
away from the storage is matched by the vertical flux of seepage through the floor of the storage.
In most cases this equilibrium will take many years to develop and may not be achieved during
the life of the storage.
After cessation of tailings disposal, seepage flux through the floor of the storage will begin to
diminish as excess water (above field capacity) drains from the deposited tailings. As the flux
through the floor diminishes below the rate of supply necessary to maintain the groundwater
mound, the balance would tum and the groundwater mound and hence rate of migration of seepage
water away from the storage would diminish.
Case2: Tailings Storage in Humid Valley Fill Sites
Tailings storages constructed in valleys are typified by relatively low floor area and high deposition
rates, particularly in the early stages of development. The tailings are confined by usually a
single main embankment with perhaps an upslope embankment and runoff diversion to limit
surface inflows. Groundwater flow is often dominated by preferred pathways associated with
geological structure. In areas of high topographic relief, the groundwater system tends to mirror
the surface topography with groundwater movement being toward creek lines. Creeks often
represent groundwater discharge areas.
Runoff is usually rapid and concentrated in a well developed drainage network. Relatively high
rainfall and lower potential evaporation mean that runoff tends to be a significant component
in the overall tailings water balance. In particular surface runoff can accumulate in the storage
over months of high rainfall and low potential evaporation.
The movement of water through the floor of a storage tends to be controlled by the permeability
of the foundation material. Along creek lines where groundwater levels are high seepage movement
can be rapid after intersecting the groundwater. Local pore pressure relief induced in the tailings
preferential seepage areas would result in generation of lateral movement of tailings water from
the surrounding tailings mass. The downgradient migration of groundwater along preferred
pathways is accelerated by imposed hydraulic pressures of the tailings. This results in potential
for increased baseflow containing tailings water in the creek(s) immediately downslope of the
tailings storage.
GILBERT
124
Water Flow in Mine Wasta Dumps.
5.0 CONCEPTUAL MODEL OF WATER FLOW THROUGH MINE WASTE DUMPS
The inherent physical characteristics of mine waste results in waste dumps having high porosity
and relatively large void spaces. This in turn results in high permeability and maintenance of
unsaturated conditions through the bulk of the dump.
The high percentage of rainfall water which does infiltrate through an uncovered dump drains
predominantly through macro-pores which form a loosely linked network of channels akin to
a highly fractured heterogenous aquifer.
In effect the bulk of the drainage down through a waste dump occurs through a relatively small
percentage of the total voids with the disperse infiltration of water at the surface concentrating
into a decreasing number of macro-pores as it moves down through the dump.
A commonly employed strategy for rehabilitation of mine waste dumps is to reduce infiltration
of water by constructing a low permeability cover (usually compacted clay) over the dump and
providing a soil/vegetation system over the top of the cover. The intention of this being to reduce
to rate of water movement into the waste pile by providing a low conductivity layer over it and
to encourage high interception and evapotranspiration losses through the provision of a soil
vegetation system at the surface. This strategy is often supplemented by construction of an
efficient surface drainage system which results in relatively high runoff and reduced opportunity
for infiltration into the dump.
Figure 1 illustrates dominate flow processes associated with this conceptual model of water
movement through waste dumps. The pertinent points to note are:
(1)
(2)
(3)
(4)
(5)
The rate of infiltration becomes limited by the rate of movement through the compacted
clay layer. Depending on design and placement quality, the vertical permeability and
hydraulic conductivity of a clay layer can vary by an order of magnitude. Settlement or
shrinkage cracking can result in local high infiltration capacity in the cover. Adequate
specification and placement quality control is required to prevent this being a problem.
In well designed and constructed rehabilitation works, high surface ponding is minimised.
Large ponded areas on the surface of waste dumps act as reservoirs for water infiltration
after storm events and are thus are undesirable particularly where they connect with high
infiltration pathways through the clay cover.
Inappropriate vegetation comprise plant with pervasive root development can lead to
development of local pathways through the soil covers.
Once through the covers, the downward migration of water through the dump would be
dominated by flow through the same macro-pore system as occurs in an uncovered dump.
The lower volume of infiltration would result in percolation through a lesser number of
macro-pores and preference for those which coincide with areas where defects in the
cover occurred.
The oxidation products which form around the waste materials are transported out of the
dump by water movement past them. This movement is limited by the quality of the cover
system and the distribution of defects and macro-pores. In effect the migration of
contaminants can be limited by limiting the opportunity for flow to occur through the macro-
pores by limiting cover defects. The resulting migration of containments being a function
of the number of active macro-pores rather than the rate of oxidation in the dump.
125 GILBERT
water Flow in Mine Wasta Dumps.








GILBERT
126
"'
a.
E
:::.
"0
Q)
en
~
.s::.
Ol
:::.
e
5
c
Q)
E
Q)
>
0
E
....
2
"' ;:
0
Qj
"0
0
E
"'
:::.
0.
Q)
u
c
0
u
Q)


- ~
"0
Q)
iii
u
0
"'
"'
"'
"'
Q)
"'
"'
Q)
(.)
0
0..
;:
g
c
"'
c
.E
0
"0
0
c
0
-

en
2
en
u:::
Second Australian Acid Mine Ora1nage Workshop (Eds N.J. Grunden and L.C. Bell}, pp 127-145
WETLANDS TO TREAT AMD- FACTS AND FALLACIES
D.R. Jones and B.M. Chapman
CSIRO Minesite Rehabilitation Research Program,
Division of Coal and Energy Technology, P.O. Box 136, North Ryde, NSW 2113.
ABSTRACT
Wetlands have many attractions for the treatment of mine drainage since they have the potential
to provide an aesthetically attractive low cost, low maintenance, and sustainable alternative
to expensive chemical treatment plants. The need for self sustaining systems becomes much
more critical following site decommissioning since not only will there be few, if any, staff on site
but there will no longer be a direct cash flow to support high operating costs. The objective
of this paper is to synthesise what is known about the fundamental physical, chemical, and biological
processes in wetlands to provide the basis for designing effective and sustainable systems for
the treatment of AMO. The roles, functions, and limitations of different components (for example,
anoxic limestone drains, surface and sub-surface flow wetlands, ponds, and riffle zones) of a
wetland treatment system are highlighted and discussed.
1.0 INTRODUCTION
Wetlands are a complex assemblage of abiotic chemical processes, aerobic and anaerobic bacteria,
emergent rooted plants, floating plants, epiphytes on the surfaces of plants, and algae co-existing
in a functional relationship. They exist wherever the underlying soil is permanently saturated.
Constructed wetlands can range from a marsh or pond created in a natural setting where one
did not exist permanently before to formed structures involving earth moving and erection of
permeable bunds and impermeable containment barriers.
Although research on wetlands has been conducted over the past 50 years, it is only over the
past decade that major fundamental advances have been made in designs for the treatment
of acid mine drainage (AMD). This has been the result of multidisciplinary collaboration between
hydrologists, chemists, and microbiologists. The primary impetus for much of this work has
been provided by the need to find economic solutions for the treatment of contaminated water
draining from abandoned and operating coal and base metal mines. These waters are often
acidic and contain high concentrations of metals (Chapman et al. 1983; Wildeman and Laudon
1989). There are several reasons why wetlands are perceived to be attractive options for treating
mine water:
(1) They are potentially low maintenance with low requirements for energy and material
input compared with chemical treatment plants;
(2) They incorporate a range of physical, chemical and biological processes which
can reduce metal concentrations to very low levels;
(3) They can be aesthetically attractive with consequent 'green' appeal; and
(4) There is the potential for recovery of metal values from the wetland substrate.
127 JONES & CHAPMAN
Wetlands to Treat AMD.
2.0 PHYSICAL CONSTRAINTS ON WETLANDS
The most important considerations governing the selection and location of a site for a constructed
wetland are:
(1) The availability of enough land with a suitable topography to provide the hydraulic
head needed to maintain passive flow through the system;
(2) The absence of large water inflows during storm events;
(3) Sufficient year-round supply of water to ensure that the wetland remains in a
permanently saturated condition; and
(4) Potential for impact on groundwater quality.
A satisfactory hydrological regime is a critical pre-requisite for the successful implementation
of a sustainable wetland. This aspect must be addressed early in the design phase to determine
if, on this ground alone, a wetland treatment system will be viable.
The drainage lines of natural catchments are prime candidates for the location of wetlands
since the requirements for earthmoving works are minimised. However, an analysis of the
catchment runoff yield should be done to ensure that the wetland is not likely to be swept away
by a storm event. If the risk of this is high then a diversion structure, or a flow equalisation
pond, must be built. Further analysis of the distribution of rain through the year is needed to
determine if a supplementary water supply (such as a dam} will be required to maintain flow
through the system. This is likely to be the case for a mine located in a tropical monsoonal
climate, or a semi-arid region where evaporation exceeds rainfall for most of the year. In the
case of a mine which is still operating, this water could be supplied by dewatering bores. If
there is potential for contamination of a groundwater resource by downward percolation of partially
treated AMD, then the bottoms of the wetland cells and interconnecting flow channels should
be sealed with plastic or clay liners.
3.0 PERFORMANCE CRITERIA FOR WETLANDS
The maximum concentrations of residual metals permitted in treated AMD will be governed
by the locally applicable water quality criteria for discharge to surface and/or groundwaters.
These criteria may be based on end-of-pipe concentrations, or concentrations in the receiving
waterway downstream of a specified mixing and dilution zone. The target discharge criteria
must be specified before a wetland can be designed since they will impact critically on not only
the size of the system, but also on the range of different types of treatment cells that may need
to be incorporated. The ANZECC water quality criteria for irrigation, stock watering, and protection
of ecosystem health are most likely to be used to specify performance targets in Australia.
4.0 PHYSICAL, CHEMICAL AND MICROBIOLOGICAL PROCESSES IN WETLANDS
Since the oxidation of iron disulfide minerals (pyrite and marcasite) is the primary source of
the acidity, dissolved forms of iron (Fe
2
and Fe', ferrous and ferric ions respectively) are usually
a major component of AMD. The composition of the parent mineral assemblage determines
the levels of other metals in AMD. In addition to Fe
2
and Fe', and sulfuric acid, high
concentrations of potentially toxic ions such as AI', Cu
2
, Cd
2
+, Zn
2
, Pb
2
, and Aso,' may also
be present (Table 1 ). The data in Table 1 show that the concentrations, and relative levels
of metals vary widely from one minesite to another in Australia. AMD from coal mines typically
JONES & CHAPMAN
128
Wetlands to Treat AM D.
contains much lower concentrations of toxic heavy metals than base metal or gold mines. In
this context it is important to note that most of the 'success stories' reported for the use of full
scale wetlands have been for systems treating such low strength and relatively 'benign' AMD.
Table 1.
Composition of AMD from different mines In Australia.
Analyte
Abandoned Gold Mine
Mt Morgan
Coal Mine Coal Mine
Mine (NSW)
(North Aust.) (Old)
(NSW) (Old)
pH
2.6
3.1
3.72
3.1 3.14
Conductivity (JlS em')
2000
6000 13600 No 8644
Sodium (mg L"')
1.61
229 213
31.7 1168
Potassium (mg L')
0.7
61.4
ND
NO 22
Magnesium (mg L"
1
)
2506
700 1700
NO 36.7
Calcium (mg L ')
12.4
268 437
ND 526
Sulfate (mg L"')
1160 5880
16100
1340 3700
Iron (mg L')
212
20.0 1460
12.6 12.1
Aluminium (mg L"')
34.6
179 803
18.1 10.3
Manganese (mg L"
1
)
1.98
55.6 131
3.57 7.9
Copper (mg L"
1
)
8.89
6.62 106
0.06 < 0.05
Zinc (mg L')
150
187 45
0065 < 0.05
Cadmium (mg L"
1
)
0.78 0.73
0.36
< 0.02 < 0.02
Lead (mg L"')
0.62
0.67
< 0.02
< 0.02 < 0002
Arsenic (mg L"')
4.72
0.13
< 0.05
< 0.005 < 0.05
The composition of the drainage is the single most important factor which will determine the
size and sustainability of a wetland treatment system. To appreciate why this is so, it is necessary
to consider the nature of the chemical and biological processes which remove acidity and metals
from AMD.
The processes that can reduce the concentrations of metals in wetlands are summarised in
Table 2. It should be noted that, although the physical processes of dilution and dispersion
act to reduce the concentration of a pollutant downstream from a source, they do not per se
decrease the total load or flux of the metal being transported.
Since Fe
2
, Fe', and Al
3
+ ions are the most abundant non-alkali and alkaline earth cations in
AMD, the chemical reactions of these species are crucial to the initial fate of the metals in AMD.
The Fe
2
oxidises to Fe' (the rate of which depends on the prevailing pH and the extent of bacterial
activity), which subsequently hydrolyses and precipitates. The rate of oxidation of Fe2 increases
exponentially above pH 5 in oxygenated water.
Depending on the flow regime, and rate of pH increase resulting from neutralisation of the
acidity, 'ferric hydroxide' can be deposited by concretionary growth on an existing coating of
the material adhering to the bed of the wetland, or precipitated via rapid nucleation. In the latter
case it forms a low density floc which must be trapped in a pond or a dense reed bed. The
aluminium ion, Al
3
+, behaves similarly but precipitation occurs at a higher pH (see Figo 1).
129
JONES & CHAPMAN
Wetlands to Treat AMD.
Table 2.
Processes In wetlands that reduce metal concentrations and acidity.
Process
Dilution
Dispersion
Oxidation
Precipitation/co-precipitation
Adsorption on precipitates
Adsorption/ion exchange on
suspended and bed sediments
SuHate reduction
Uptake by biofilms, algae, and
aquatic macrophytes
Nature Controlling Variables
Physical Volume of Water
Physical Flow velocity, channel geometry and
roughness.
Chemical Concentrations of oxygen and organic carbon
Microbiological
Chemical
Chemical
Physical
Chemical
Microbiological
Chemical
Biological
Concentration of components, pH, redox
potential
Amount of precipitate, concentrations of major
cations and anions, pH
Concentration of suspended sediment, particle
size and mineralogy, pH, concentrations of
major cations and anions
pH, organic carbon, concentrations of oxygen
and sulfate
Density of plants, temperature, light intensity,
availability of nutrients
100
- --------;;;----
;'
~
;'
"
I
80
co
Q) 60
2
cf.. 40
20
Fig. 1.
0
2
I
AI(OH)3
2+ 2+
Fe(OH)3
I
Ni Cd
I
,
I
** *
I 2+
/*
Pb
2+
I
I
2+
Zn
I
Cu
I
/
I
I
I
I
I
I' /
/
..... / !- ..
-
3 4 5 6
pH
Schematic showing precipitation of Fe(OH)
3
and AI(OH)
3
, and pH
50
values
for adsorption of Pb
2
+, Cu
2
', Zn
2
', Nl
2
' and Cd
2
' on Fe(OH)
3

The chemical equations for the oxidation and subsequent hydrolysis and precipitation of Fe
3
+,
and for the precipitation AI"' are depicted in Eqn. [1] to Eqn. [3]. The precipitation of the hydroxides
generates acidity. This incipient source of acid must be taken into consideration when calculating
the total acid load to be treated by a wetland. A pH value of at least 6 should be achieved prior
JONES & CHAPMAN
130
Wetlands to Treat AMD.
to moderate strength AMD entering a surface flow wetland. If the pH is less than 6 there will
be insufficient alkalinity present to prevent the pH falling substantially as a result of the oxidation
and hydrolysis of Fe''.
Fe'' + H' + _!_o, -t Fe
3
' + _!_H,O
4 2
Fe
3
' + 3H
2
0 --t Fe(OH)
3
+ 3H'
A/
3
' + 3H,O --t A/(OH)
3
+ 3H'
[1]
[2]
[3]
Under oxidising conditions, and for the pH range 4 to 7, the dissolved concentrations of trace
levels of Cd, Cu, Ni, Pb, and Zn in mine water are likely to be limited by adsorption on precipitated
ferric and aluminium hydroxides. Indeed the high scavenging capacity of these hydroxide forms
the basis of most industrial chemical treatment processes for the removal of heavy metals.
Adsorption occurs over a typically fairly narrow range of one to two pH units and is manifested
by a large decrease in the concentration of metal in solution as the pH increases. The pH range
over which this decrease in concentration occurs is called the pH-adsorption edge, and the pH
at which the dissolved metal ion concentration declines to 50% of its initial value is termed the
pH
50
value. The pH50 value for a given metal is a complex function of the nature of the sorbing
phase (the sorbent), the ratio of the concentrations of sorbate (the metal ion) to sorbent, and
the concentrations of major cations, anions, and dissolved organic matter. For crystalline goethite
(an iron oxide which is typically found in AMD systems) the pH
50
values for Pb, Cu, Zn, Ni and
Cd are 3.2, 4.3, 5.5, 5.7 and 6.0 respectively for initial metal concentrations of 20 J.Lg L-1 (Fig. 1 ).
As the initial metal concentration to iron oxide ratio increases the pH
50
value also increases.
Thus a higher pH will need to be achieved before 50% of the metal is removed from solution.
This fact has important implications for AMD which contains high concentrations of heavy metals
relative to the initial concentration of dissolved iron.
The pH of AMD should be raised to at least pH 6 prior to the oxidation of Fe' to Fe3' to make
the best use of the metal adsorbing capacity of the ferric hydroxide precipitate.
For pH values greater than 7, the solubilities of metal cations present at trace levels can also
be limited by the precipitation of hydroxide/oxide and carbonate compounds. However, It should
be noted that the solubilities of amphoteric metals such as Zn and Pb decline to a minimum
with increasing pH, but then start to increase again at higher pH values as a result of the formation
of more soluble negatively-charged hydroxy complexes. Thus too high a pH can be detrimental
for the removal of these metals.
Metals can also be removed from solution by adsorption on the surfaces of biofilms, on algae,
or by epiphytes growing on the surfaces of aquatic plants. This route may be very significant
if there is extensive in--stream vegetation.
Under reducing conditions, in the presence of sulfide ion, the solubilities of Cu2', Pb2', Zn2',
Ni
2
', and Cd
2
' will be limited by the precipitation of the respective insoluble metallic sulfide
compounds or by co-precipitation with ferrous sulfide. The sulfide salts of most heavy metals
are much less soluble than their hydroxide and carbonate counterparts within the normal pH
range for soil environments (Patterson 1985). This is particularly important for Cd'' and Ni'
where the pH would have to be raised above 10 in order for discharge quality criteria to be
satisfied by the precipitation of the metal hydroxides.
131
JONES & CHAPMAN
Wetlands to Treat AMD.
Recent research has highlighted the critical importance of facilitating reduction of sulfate in wetlands
in order to ensure indefinite operation at high levels of elliciency (Hedin et al. 1989; Brierley
1990; Mcintire et al. 1990; Dvorak et al. 1992; Eger 1992; Hammack and Edenbom 1992; Wildeman
et al. 1992). Sulfate reduction rates have been measured in marine and freshwater sediments
and have ranged from 0.038 ~ g cm-
3
d-
1
to 290 ~ g cm-
3
d-
1
Rates can vary over several orders
of magnitude across a pond or lake owing to heterogeneity in redox status and the amounts
of organic matter present in the sediments. Fine-grained sediments are likely to exhibit higher
rates owing to reduced penetration of oxygenated water from the overtying water column). Bacterial
sulfate reduction rates have been measured in a constructed wetland containing mushroom
compost as a carbon source and were found to range from 0.19 ~ g em-" d-
1
to 58 ~ g em-" d-
1
(Mcintire et al. 1990).
Sulfate reducing bacteria represented by the genera Desulphovibrio and Desulphatomaculum
are obligate anaerobes that require oxygen concentrations to be less than 0.16 mg L -
1
Their
pH optimum lies between 5 and 9, and, with acclimation, they can tolerate temperatures from
0 oc to 50 oc (Gioyna and Espino 1969; Postgate 1979). Since the lower concentration limit
of sulfate to support their activity is approximately 30 mg L-
1
, sulfate limitation will not be an
issue for the treatment of even very dilute AMD. The reduction of sulfate is coupled to the oxidation
of organic compounds to bicarbonate:
sot + 2 CH
2
0 + 2W ---> H
2
S + 2H
2
0 + 2C0
2
sot+ 2CH,Q + OH----> Hs- + Hp + 2HCO;
(pH< 7.0)
(pH> 7.0)
This alkalinity provides additional neutralising capacity in the system and will help to buffer it
against transient shock loadings of increased acidity. There is generally no shortage of sulfate
in drainage from minesites. However, these waters are generally deficient both in nutrients
(P and N) and in organic carbon that is the 'fuel' for sulfate reducing bacteria. Consequently,
it has been found that readily degradable organic matter such as compost needs to be added
to the gravel substratum of a wetland in order to initiate the formation of an anoxic zone (Mcintire
et al. 1990; Hedin and Nairn 1992).
Sustainable removal and retention of sulfate in wetlands depends on the maintenance of anaerobic,
strongly reducing conditions. Wetlands constructed for the removal of metals as sulfides must
be designed to maintain these conditions and avoid exposure of the sediments to oxidising
conditions.
Manganese (Mn
2
) has proved to be one of the most difficult metals to remove from mine drainage.
It does not readily form sulfide minerals and will not precipitate as a carbonate or hydroxide
until relatively high pH values are attained. It has been found that the most effective removal
occurs above pH 7 to 8 when a combination of abiotic and biotic processes are operative (Chapman
et al. 1988). The Mn
2
is oxidised and precipitated as a Mn or a mixed Mn
3
/Mn hydrous
oxide. This oxidation can be promoted by bacteria (Batal et al. 1989). However, once a thin
film of Mn-oxide has been deposited, an abiotic autocatalytic reaction will proceed readily if the
pH is above 8.
In summary, pH and the presence or absence of oxygen are the key parameters governing
the removal of metals and acidity from AMD. The pH must be higher than 6 in order for there
to be effective treatment by aerobic or anaerobic processes. If sulfate reduction is a major
component of the treatment process then the wetland must not be allowed to dry out or else
the sulfides will re-oxidise to produce acid, and remobilise the initially trapped metals.
JONES & CHAPMAN 132
Wetlands to Treat AMD.
5.0 OPERATIONAL COMPONENTS OF WETLAND TREATMENT SYSTEMS
A fully functional and self-contained constructed wetland treatment system must, of necessity,
contain a number of process units in addition to a series of linked ponds containing plants.
These include pre-neutralisation systems (active and passive chemical), aeration zones, at least
one of four different types of 'wetland' cell designs, and algal-filters. The functional role of each
of these components, and the way in which they should be incorporated into a wetland treatment
system is explained below.
5.1 Pre-Neutralisation of AMD
As stated above, pre-neutralisation of AMD is the key to efficient removal of metals by aerobic
or anaerobic wetland treatment cells. Ideally the pH should be raised to 6 before the effluent
contacts the more sensitive biological components of the wetland proper.
5.1.1 Active chemical
For highly concentrated AMD (such as occurs at many Australian mines), there may be no
alternative but to pre-treat the initial drainage with lime or magnesia. Sodium hydroxide is generally
prohibitively expensive for this purpose. Unlike chemical treatment plants, where the final pH
may need to be as high as 10, to completely remove the metals, the target pH for pre-neutralisation
for a wetland treatment system only needs to be about 6. The reagent should be added to the
effluent prior to a cascade system to promote oxygenation of the treated stream.
5.1.2 Passive chemical (Anoxic Limestone Drains)
The installation of limestone-containing drains prior to a wetland can help to overcome the problems
caused by low pH water (Nairn et al. 1991; Nairn et al. 1992). Experience has shown that an
anoxic environment should be created in these drains. Anoxic conditions prevent the oxidation
of Fe
2
to Fe
3
, which will passivate limestone under aerobic conditions by coating it with a layer
of ferric hydroxide precipitate. Ferrous hydroxide and ferrous carbonate are much more soluble
than ferric hydroxide at pH 7 to 8, and thus Fe
2
is unlikely to precipitate in an anoxic drain.
Anoxic limestone drains (ALDs) are relatively simple to construct (Fig. 2). They consist of a
trench filled with limestone, which is covered with a layer of plastic and low permeability soil
or clay to prevent the ingress of oxygen. It is recommended that high purity limestone (> 90%)
be used for the construction of the drains since the rate of alkalinity generation (dissolution of
the carbonate) is greatest for this material. The bed should be flooded with water prior to the
installation of the capping layer to minimise the volume of air entrained within the drain. The
initial development of anoxic conditions could be accelerated by the provision of a good carbon
source for bacteria. The drain is ready to receive AMD once the dissolved oxygen concentration
drops to near zero. In order to minimise the ingress of oxygen into the drain, the inflow and
outflow points should be protected by submerging them in sumps. It is also very important to
locate the inflow for the drain as close as possible to the source of the AMD to limit the extent
of oxidation of Fe
2
once the AMD contacts the atmosphere.
The critical parameters which ultimately determine whether an anoxic limestone drain will be
viable are the concentrations of Fe
3
and AI' in the source water, since both of these rnetal
ions will form insoluble precipitates at the prevailing pH. Aluminium hydroxide is especially
troublesome since it forms a gelatinous mass which can rapidly block the flow paths through
the drain. Based on US experience, the concentrations of Fe' and AI' should each be less
than 2 mg L-' to ensure long term performance. Reference to Table 1 shows that, according
133
JONES & CHAPMAN
Wetlands to Treat AMD.
Plastic Membrane
/ / 7
I
77/Fr
'/ / / / / / / / / / / / / / /
t Outflow
Inflow
/ / / / / /

/ / / / /

/ / / / / / / / / / / / / / / /-
Crushed Limestone
+ 2+
CaC0
3
+ 2H c::=:;> Ca + H
2
C0
3
CaC0
3
+ H
2
C0
3
c::=:;> Ca
2
+ + 2HC0
3
-
Fig. 2. Cross section through an anoxic ffmestone drain.
to these criteria, none of the examples given would be suitable for treatment by an ALD. In
practice, however, lifetime which is somewhat less than the ideal might be an acceptable
compromise.
Since ALD technology is only about five years old, there are no data on long term performance.
The effects of armouring reactions and changes in permeability of the bed are unknown. This
means that design criteria are still very much in the formative stages. However, the minimum
volume of limestone required for a given lifetime can be calculated using, as the starting point,
the total acidity (proton plus metal) expected to pass through the system over this period. On
this basis, it is recommended (Hedin et al. 1994) that, for a design life of 30 years, an ALD should
contain about 8 t limestone L-' min-' of AMD containing 400 mg L-' acidity (CaC0
3
equivalent).
In order to realise the full potential of an ALD for the generation of alkalinity, the residence time
of water in the system should be about 14 h. Actual performance will depend critically on mass
transfer limitations which will be a function of flow rate, geometry, and the rate of armouring.
5.2 Types of Wetland Treatment Units
Four basic types of wetland treatment units can be identified:
(1)
(2)
(3)
Free water surface (FWS) systems which have predominantly surface flow with
shallow water depths and extensive growths of emergent aquatic plants throughout;
Subsurface flow (SSF) systems in which most of the water flows laterally through
a bed of sand or gravel, which may be planted with emergent aquatic plants;
Subsurface flow (SSF) systems in which the water flows vertically upwards or
downwards through a permeable substratum, which does not contain plants; and
JONES & CHAPMAN 134
Wetlands to Treat AMD.
(4) Lagoons- ponds several metres deep with floating plants in middle of basin and
rooted emergent plants around periphery. Lagoons serve primarily as sedimentation
basins. However, if sufficient organic matter is present in the bottom sediments,
microbial respiration can lead to anaerobic conditions which favour the
immobilisation of many elements.
Each of the wetland cell designs is discussed below and the scale at which operational experience
has been gained is indicated in italics.
5.2.1 Free water surface wetlands (full scale operational)
The first generation of constructed wetlands were of the surface flow type in which predominantly
aerobic zones were in contact with the water (Fig. 3). In these systems adsorption of metals
by hydrous oxides of Fe and Mn and surface organic matter, and precipitation as hydroxide
or carbonate salts are the primary removal mechanisms. Many of these wetlands have been
installed in the United States to treat low strength AMD from coal mines. The most important
operational requirement for these systems is that the pH of the water flowing into them must
be near 6, or there will be insufficient removal of the metal ions by precipitation or adsorption
to prevent the ultimate exhaustion of binding sites in the substrate. This breakdown in aerobic
surface flow wetlands is especially likely to occur if the input water has a pH lower than about
6. At lower values than this there will be insufficient bicarbonate alkalinity to buffer the acidity
generated by the precipitation of ferric hydroxide which is the dominant metal ion-adsorbing
solid in this type of wetland (Brodie 1990).
Thus pre-neutralised AMD and anoxic seepage water, at pH 7 to 8, are the best input sources
for FWS wetlands since such waters contain an abundant supply of iron, which will precipitate
as ferric hydroxide and bind the metals. Although the substratum a short distance beneath the
soil-water interface in a FWS wetland may become anaerobic, removal of solutes via the
precipitation of sulfides will be strongly limited by diffusion. Hence longer retention times are
required for these systems to make effective use of this removal mechanism than for subsurface
flow wetlands. For related reasons, the addition of limestone to the substrate of a FWS wetland
is unlikely to be of much benefit. A similar situation pertains in a pond or lagoon.
Surface flow wetlands most closely resemble natural systems and are likely to be the most robust
and require the least long term maintenance of any of the options for wetland cell design, provided
that the inflow water is at the correct pH. Their one major disadvantage is that they are less
efficient, and hence require a larger area, than the subsurface designs discussed below.
5.2.2 Subsurface lateral flow (pilot scale experimental)
In contrast to FWS wetlands, SSF systems are more costly to construct, and rigorous control
of substrate type and placement is required. However, it is now recognised that SSF wetlands
can be much more compact than their FWS counterparts, for a given throughput, by virtue of
the higher surface area to volume ratio presented to the mine water. In lateral flow SSF wetlands
the water flows horizontally through a gravel bed, after being introduced via a distribution system.
The design may include provision for a shallow layer of water on top of the substrate.
SSF wetlands are particularly attractive tor those cases where sulfate reduction is to be used
for the removal of heavy metals since anaerobic conditions can readily be created by the addition
of a supply of organic matter (such as compost) to the substrate. Wetlands of this type which
have been supplemented with added organic matter have been referred to in the literature as
'compost' wetlands. Unlike FWS wetlands, the addition of limestone to the substrate will be
135 JONES & CHAPMAN
Wetlands to Treat AMD.
Fig. 3.
Free Water Surface (FWS)
(Aerobic)
( +) Low maintenance (-) Area
Requirement - pH>6
Subsurface Flow (SSF)
(Aerobic/ Anaerobic)
Lateral
\D;_:& fX:Jc tj,+
..... . . . . . .
- . .. :.
Vertical , ,
downflow "j<Sdf:ififJ.O 66J?J;f
.. ... CI:.
1ft,fVflf!RrliJiiN
( +) Sulfate reduction
(+) Less area than FWS or
lateral flow SSF
(-) Blockage of
flow paths
(-) Exper"tmental
Requirement - pH - 5
Schematic diagrams of four different types of constructed wetland cells.
JONES & CHAPMAN 136
Wetlands to Treat AMO.
very beneficial owing to the much greater contact area, and the reducing conditions which prevent
armouring by ferric hydroxide (c.f. anoxic limestone drains). The limestone may help to protect
the wetland from damage if a pulse of much poorer quality water than usual enters the cell,
and is likely to be especially beneficial during the initial establishment phase when the substrate
is being colonised by sulfate reducing bacteria.
Although organic matter must initially be added to the substrate, the long term supply of decaying
organic matter can be assured by growing aquatic plants in the substratum. One of the major
unanswered questions about long term performance relates to the blockage of flow paths by
plant roots, precipitates, or bacterial biomass, with consequent reduction in performance. For
this reason SSF wetland cells should be located in series after a FWS cell to minimise the loading
of particulates in the inflow.
5.2.3 Subsurface vertical flow (small scale experlmentaf)
Recent research has indicated that wetland cells for the removal of sulfate and toxic metals
from AMD should be designed as upflow units for maximum removal efficiency/unit area. In
these systems the water is injected at the bottom and percolates upwards through an anaerobic
zone (Morea et al. 1989; Wildeman et al. 1992). Crushed limestone can be added in a layer
at the bottom of an upflow cell to initially supplement the acid neutralising capacity of the cell
in the period before sulfate reducing activity is fully established.
Although water can be transmitted horizontally through a bed of organic substrate (as in the
subsurface lateral flow systems discussed above), the hydraulic head is much larger and is
more easily controlled in a vertical system. A vertical flow unit can be designed with a ten times
higher flow rate than an identically sized horizontal system (Eger 1992). The maximum flow
rate that can be used in practice before there is an unacceptable deterioration in performance
will be determined by the balance between residence time and the rate of removal of the metal
ions in the bed. However, the long term effective operation of a vertical flow unit will depend
even more critically on the rate of occlusion of the flow paths. For this reason, vertical flow
cells should be located downstream of FWS cells to minimise the concentration of particulates
in the inflow. It is recommended that plants not be grown in the substrate of upflow cells
Consequently, periodic additions of readily degradable organic matter will need to be made to
these cells to sustain the population of sulfate reducing bacteria. Thus, the advantages of smaller
size and higher efficiency of upflow cells is offset by higher maintenance requirements.
5.2.4 Ponds (Lagoons)
The main role of ponds (up to 2m deep) in wetland systems is as sedimentation basins to trap
precipitates of ferric and aluminium hydroxides generated from neutralisation and oxidation reactions
occurring further upstream. This sedimentation and trapping function is typically enhanced with
stands of Typha and Eleocharis. Organic matter, either added or produced over time by the
decay of plants. on the bed of the pond will facilitate sulfate reduction. Ponds should precede
FWS or SSF wetland cells to reduce the loading of solids into these components of the system.
Low solids loading is especially critical for the long term functioning of SSF systems. Whilst
ponds are simple to construct, and are vital components of a wetland treatment system, thermal
stratification can considerably reduce operating efficiency. If stratification occurs, a thin layer
of water moving at relatively high velocity can transport untreated water across the surface of
the pond to the discharge point. Thermal stratification is especially likely to occur when there
is a large diurnal range in air temperatures. To prevent a stratified layer from forming, a curtain
hanging from a floating boom can be installed near the upstream end of a pond to force the
incoming surface water to mix with the lower layers.
137
JONES & CHAPMAN
Wetlands to Treat AMD.
5.3 Riffle Zones and Algal Mats
Effective oxygenation of AMD is critical lor the rapid oxidation of Fe
2
to Fe
3
' at the beginning
of the wetland treatment system, and the oxidation of Mn
2
' to Mn'' in the final polishing section.
The dissolution of oxygen is enhanced by creating turbulence in the water. This can be achieved
by linking a series of cascades with a shallow channels lined with coarse rock (rip rap). The
length of the cascade that is required will depend on the input concentration of Fe
2
'. The pH
of the water entering the cascade should be near 6 to not only maximise the rate of oxidation
but to make maximum use of the adsorptive capacity of the ferric hydroxide that is produced.
Growth of algae in the riffle zone will further enhance the process of precipitation and metal
removal.
Riffle zones should also be located after subsurface flow cells to oxygenate the water and to
provide optimum conditions for the removal of Mn
2
'. The pH of the water leaving the final wetland
cells should lie between 7 and 8. At this pH, algae can very effectively oxidise Mn
2
' to Mn
4
',
which will precipitate. Manganese is regarded as the most difficult metal to remove from AMD.
Indeed some overseas authors have recommended that the size of an FWS wetland for the
removal of Mn should be three times larger than that required for the removal of an equivalent
amount of iron. However, the specific conditions needed for the efficient removal of manganese
have not been provided in most wetland systems that have been reported in the literature.
Based on work done in a passive treatment system for water from the Hilton mine (Chapman
et al. 1988), the authors of this paper found that Mn
2
' can be very efficiently removed from pH
neutral water. Filamentous blue green algae (cyanobacteria) attached to the bed of a shallow
stream created a high pH (> 9) oxygen-rich microenvironment during the day, and Mn oxides
precipitated on the surface oJ the algal cells. The work of the algae was enhanced by an abiotic
process of autocatalytic oxidation of Mn
2
' that occurred on the surfaces of the gravel on the
bed of the stream.
6.0 ROLE OF EMERGENT PLANTS IN WETLANDS
The role of plants in wetlands has been somewhat controversial. Detailed studies have shown
that, in contrast to the floating plants such as duckweed and water hyacinths used in lagoon-based
systems, uptake of metals into the tissues of rooted emergent aquatic plants (e.g. Typha) in
wetlands is not a major route for metal removal from mine waters containing higher than trace
level concentrations of metals. It has been estimated that less than 5% of the metals input into
a wetland bed are incorporated into the plants (Sencindiver and Bhumbla 1988; Machemer et
al. 1990). Plants do, however, enhance the removal of solutes by indirect means. In the case
of FWS systems plants reduce water velocity and lead to the sedimentation and trapping of
metal enriched particulates, and provide a substrate for epiphytic algae. The reduced velocity
also increases the residence time and lengthens the treatment period for a given parcel of water.
The second important role of plants is as a long term source of the decaying organic matter
required to sustain sulfate reduction after the initial organic amendment in the wetland has
been consumed. However, for units designed to treat minewater at high throughput, periodic
supplementation with additional organic matter may be periodically required to maintain the
metal removal rate.
The third function of rooted emergent plants is to create oxidised microzones around their root
networks that extend into the anaerobic substratum (Brix 1987; Macfie and Crowder 1987; Dunbabin
et al. 1988). These sites promote the precipitation of high surface area iron and manganese
JONES & CHAPMAN
138
Wetlands to Treat AMO.
oxyhydroxide phases (Tarutis and Unz 1990) which have considerable scavenging capacity
for heavy metals (Jenne 1968). Consequently, metal ions that escape from the lower anaerobic
zone can be immobilised in the aerobic regions located around the roots and at the top of the
gravel bed. The microenvironment in the vicinity of roots also favours the formation of pyrite
from the iron monosulfides initially formed under strongly reducing conditions in sediments.
7.0 STATE-OF-THE-ART
Historically, the majority of wetlands constructed for the treatment of mine waters in Australia
and overseas have consisted of a combination of FWS and lagoon components, located in
existing stream channels. These systems are characterised by a predominantly aerobic environment
in contact with the flowing water. The dense stands of reeds growing in these systems serve
as a support for bacteria and algae, and act as a hydraulic barrier to reduce the speed of the
water, and enhance residence time.
7.1 Overseas
An overview of the literature reveals that most of the full-scale wetlands that have been constructed
have been treating AMD that is not only relatively dilute, but which contains relatively low
concentrations of toxic heavy metals (Wieder 1989). Moreover, the fact that the oldest functioning
wetland for treating AMD has been operating for only ten years, means that there is a lack of
long term performance data even for relatively low grade AMD. Most large wetlands built thus
far have been of the FWS type. Subsurface flow systems for the treatment of AMD are still
at the experimental or pilot scale stage.
7.2 Australia
The use of constructed wetlands for treating mine drainage in Australia has so far been quite
limited and mainly confined to situations where the effluent entering the wetland is close to neutral
pH. There is virtually no information available for experimental systems which may be in the
process of being trialled for low tci moderate strength AMD.
Examples of functioning wetlands treating near-neutral pH mine water are located at the Ranger
uranium mine (NT), Tom's Gully gold mine (NT), Woodcutters base metal mine (NT), the Hilton
base metal mine at Mt lsa (Old), the Hellyer mine (Tas) and the Westralian Sands Ltd synthetic
rutile plant (WA).
At the Ranger mine, the solutes of interest are uranium (in the form of UO{'), Mg2, SO/, and
Mn
2
' (D.R. Jones unpublished data). The constructed wetland consists of three shallow ponds,
(each 1.5 m deep and 20 x 20m in area) in series which discharge into a retention basin. Results
so far indicate effective removal of U at flow rates up to 1 ,000 m' d-1
The Tom's Gully gold mine discharges drainage into an ox-bow billabong containing Typha (Noller
et al. 1992). This overflows into the Mount Bundy Creek. A range of transition and base metals
as well as As are effectively removed in this system. This is an example of a natural wetland
in the vicinity of a mine that can be used to improve the water quality of mine drainage.
At Woodcutters mine, effluent with near-neutral pH passes through a settling pond and then
discharges to a pond and wetland channel, prior to being released into a creek. Metals removed
include Cd, Pb, Cu and Zn.
The water at the Hilton mine is slightly acidic as a result of being supersaturated with C0
2
139
JONES & CHAPMAN
Wetlands to Treat AMD.
(Chapman et al. 1988). However, the pH rises to over 8 after degassing. The wetland consists
of a riffle zone feeding into a 4.2 ha retention pond (average depth of 1.3 m), which overflows
into a natural creek channel. This system, populated by reeds and algae, has been studied
in detail and its efficiency optimised to an extent that the water can be reused for ore processing.
Elements being removed include Fe, Zn, Mn and Tl. The important processes that have been
identified include degassing, precipitation, coprecipitation, adsorption, sedimentation, autocatalytic
oxidation of Mn' and algal photosynthesis leading to a microenvironmental rise in pH. This
system has functioned successfully for flow rates of up to 3,000 m' d-'.
At the Hellyer mine, a large wetland consisting of a number of FWS cells (populated with Juncus
sp.) arranged in a parallel configuration was constructed in late 1992. This followed trials with
a small pilot scale system. The wetland receives pH neutral input containing Pb, Fe, Mn, and
sulfate from the tailings dam. This has been a relatively expensive system to construct.
The Westralian Sands synthetic rutile plant (Masters 1988) produces an acidic effluent which
is neutralised with lime before it is discharged into a 2 ha wetland. This system, commissioned
in 1986, was designed as a backup for the chemical treatment plant, rather than as a primary
treatment facility. The elements removed from the usually pH neutral water include Fe, Mn
and sulfate. However, following process malfunction the pH of the water can be as low as 3
to 4. Under these circumstance the wetland has been found to provide about 4 days of treatment,
before its buffering capacity is exhausted. The system has required extensive maintenance
involving the replacement of vegetation and substrate (Ryan and Hosking 1992). This is possibly
a result of the damaging effects of the acute shock loads to which it is periodically exposed.
8.0 DESIGN FACTORS AND COST
The design of a wetland will depend critically on the quantity of effluent to be treated, the types
of contaminant to be removed, the local climatic conditions, and the geomorphology and land
area available. A comprehensive overview of the many facets needed to be considered when
designing wetland systems for the treatment of wastewaters is provided in a book written by
Reed et al. (1988).
The aspect ratio most often used for wetland cells lies between 5:1 and 10:1 (L:W) to provide
maximum path length for a given area. Single or multiple cells in series or in parallel can be
used, but care has to be taken to avoid channelling or shortcircuiting. Serpentine arrangements
using channels or baffles increase contact time and mass transfer between substrate, biota
and water. In some cases, clay or synthetic liners are necessary if protection of the water table
is of concern. Depending on the requirements for the wetland, substrates such as clay, organic
soils or mushroom compost may need to be introduced and macrophyte species planted. Aesthetic
considerations can also be important in design and the wetland can eventually become a valuable
wildlife refuge. Wetlands should be designed to avoid drying out as this can dramatically reduce
the biomass, and mobilise metals initially trapped as sulfides.
As stated above, most operational experience in the use of constructed wetlands for the treatment
of AMD has been obtained with FWS systems. Not only are the design factors for this category
of wetland still largely empirical, but those factors which have been published are primarily for
the removal of Fe and AI from low strength AMD. Brodie and others have proposed guidelines
(Table 3) based on their work on FWS wetlands constructed by the Tennessee Valley Authority
(Brodie et al. 1988). These are intended as a guide o'lly and depend on many factors not
specifically considered in this table. The criteria for Mn are generally based on suboptimal
configurations tor the removal of this element.
JONES & CHAPMAN
140
Wetlands to Treat AMD.
Table 3.
Sizing guidelines for FWS constructed wetlands (from Brodie eta!. 1988).
For target discharge limits:
Influent pH < 5.5
Fe
2 m' of wetland mg-' Fe in inflow min-'
Mn 7 m' of wetland mg-' Mn in inflow min-'
Influent pH > 5.5
Fe 0.75 m' of wetland mg-' Fe in inflow min-'
Mn 2 m
2
of wetland mg-' Mn in inflow min-'
Fe
Mn
~ 3 mg L-'
< 2 mg L-'
Design Factor
2 m' mg-' min-'
7 m' mg-' min-'
0.75 m' mg-' min-'
2 m
2
mg-' min-'
There has been little quantitative evaluation of the effect of temperature on wetland performance.
This is particularly relevant for the Australian context where many mining provinces have year
round temperatures that are considerably higher than in the USA and Europe from where published
design guidelines have originated.
Costs of constructing a wetland can vary enormously and depend particularly on the topography
of existing land forms, whether natural or synthetic liners are required, and the composition of
the effluent. Typical price ranges quoted for FWS wetlands in the United States are US$8 to
US$56 m-
2
(Skousen et al. 1992), US$3 to US$32 m-' {Wieder 1989), and US$66 to US$139 m-2
(Hie! and Kerins 1988).
9.0 ASSEMBLING THE WETLAND
Based on the above discussion, a conceptual design can now be formulated lor a wetland to
treat moderate strength AMD. This design is shown in Fig. 4. The abiotic, pre-neutralisation,
oxidation, and settling components are located before the wetland components to ensure maximum
removal of acidity and metals before the more sensitive biological processes are exposed to
the drainage. Experience on pilot scale systems has indicated that FWS cells should precede
SSF cells to minimise the potential for blockage of the substrate. The latter serve to remove
the residual concentrations of heavy metals via the precipitation of sulfides. However, in particular
cases it may be necessary to install a SSF cell prior to the first FWS unit. This could occur
for the situations where the first part of the system cannot achieve a pH near 6, or where the
concentrations of heavy metals are still too high to permit the growth of higher plants. Sulfate
reduction will proceed at a pH which is lower than that needed for efficient removal of heavy
metals ions by adsorption on ferric hydroxide. This process is not only more tolerant to elevated
concentrations of these metals, but also generates alkalinity.
If the concentrations of manganese in the treated AMD are likely to exceed discharge criteria,
then a riffle zone incorporating an algal mat should be installed as the final stage of the system.
Filamentous blue green algae are preferred for this role since they can create and maintain
a higher pH in the water column during daylight hours than can green algae.
141 JONES & CHAPMAN
Wetlands to Treat AMD.

-- f-
-f-
FWS 1-
Wetland - -
- - "'
--
--
Discharge
...-Diversion Bund
Settling
Pond


:I SSF
Wetland

. ;'1/. <::::1

. ';"'[)'!'#: ..
. pH> 7
Riffle
Zone
pH >8
Fig. 4. Conceptual design of a complete wetland system for treating AMD.
JONES & CHAPMAN
142
Wetlands to Treat AMD.
10.0 LONG TERM CONSIDERATIONS
The design lifetime of a wetland treatment system is a key issue. In some cases significant
volumes of'AMD requiring treatment may only be produced during the life of the mine, prior
to rehabilitation of source material (e.g. waste rock dumps). In this case there would obviously
need to be less emphasis on long term sustainability. Wetlands, by their very nature will accumulate
heavy metals and so the long term implications of a contaminated site should be addressed
when a wetland system is being planned. In all cases performance monitoring will be required
throughout the life of the system. Typically the frequency and spatial intensity of monitoring
would need to be highest during the initial establishment and proving phase. The monitoring
program should be designed to not only measure the end-of-pipe quality but also to provide
an early warning of a decline in performance in the upstream components of the system.
11.0 CONCLUSIONS
A properly constructed wetland treatment system does have the potential to be self sustaining
with construction and operating costs that are lower than for a conventional chemical treatment
plant. The hydrological regime and the chemical composition of the drainage are the key factors
which will determine the complexity and size of the system required to produce effluent suitable
for discharge to receiving waters. It is important to recognise that most full scale systems that
have been built to date have either been treating low strength AMD or mine drainage which
has an initial pH above 4. Systems to treat moderate strength, low pH AMD are still at the
developmental stage.
Wetlands for the treatment of more concentrated sources of AMD are unlikely to be walkaway
and maintenance free. For these cases pre-neutralisation systems will require replenishment,
sedimentation ponds will need to be periodically cleaned, and the substrate in some of the more
heavily impacted upstream wetland cells may need to be replaced.
In all cases, long term monitoring of performance will be required.
12.0 REFERENCES
Batal, W., Laudon, L.S., Wildeman, T.R., and Mohdnoordin, N. (1989). Bacteriological tests
from the constructed wetland of the Big Five Tunnel, Idaho Springs, Colorado. In
'Constructed Wetlands for Wastewater Treatment; Municipal, Industrial and Agricultural'.
(Ed. D.A. Hammer.) pp. 550-557. (Lewis Publishers.)
Brierley, C.L. (1990). Bioremediation of metal-contaminated surface and groundwaters .
Geomicrobiology Journa/8, 201-223.
Brix, H. (1987). Treatment of wastewater in the rhizosphere of wetland plants-the root-zone
method. Water Science and Technology 19, 107-118.
Brodie, G.A., Hammer, D.A., and Tomljanovich, D.A. (1988). Constructed wetlands for acid
control in the Tennessee Valley. In 'Mine drainage and surface mine reclamation
(conference proceedings). Volume 1: Mine water and mine waste, Pittsburgh, PA USA,
19-21 Apr 1988'. US Department of the Interior, Bureau of Mines Information Circular
IC9183. pp. 325-331.
Brodie, G.A. (1990). Treatment of acid drainage using constructed wetlands, Experiences of
the Tennessee Valley Authority. In 'Proceedings, 1990 National Symposium on Mining'.
(Ed. D.H. Graves.) pp. 77-83. (University of Kentucky Publications: Lexington.)
Chapman, B.M., Jones, D.R., and Jung, R.F. (1983). Processes controlling metal ion attenuation
in acid mine drainage streams. Geochimica et Cosmochimica Acta 47, 1957-1973.
Chapman, B.M., Jones, D.R., Jung, R.F., Jones, C.J., Kadletz, 0., and Teague, J.W.S. (1988).
143 JONES & CHAPMAN
Wetlands to Treat AMD.
Treatment and Utilisation of Hilton Mine Water. In 'Proceedings of the 3rd Inti. Mine Water
Congress, Melbourne, Australia. pp. 147-156. (Australasian Institute of Mining and
Metallurgy.)
Dunbabin, J.S., Pokorny, J., and Bowmer, K.H. (1988). Rhizosphere oxygenation by Typha
domingensis Pers. in miniature wetland filters used for metal removal from wastewaters.
Aquatic Botanical29, 303-317.
Dvorak, D.H., Hedin, R.S., Edenborn, H.M., and Mcintire, P.E. (1992). Treatment of metal-
contaminated water using bacterial sulfate reduction: results from pilot-scale reactors.
Biotechnology and Bioengineering 40, 609-616.
Eger, P. (1992). Wetland Treatment for Trace Metal Removal from Mine Drainage: The Importance
of Aerobic and Anaerobic Processes. In 'Proc. Inti. Specialist Conference Wetlands
Downunder-Wetland Systems in Water Pollution Control, Uni. Of NSW, Sydney, NSW,
Australia, Nov. 1992'. pp. 46.1-46.9.
Gloyna, E.F., and Espino, E. (1969). Sulfide production in waste stabilization ponds. Journal
of the Sanitary Engineering Division ASCE, 95, SA3.
Hammack, R.W., and Edenborn, H.M. (1992). The removal of nickel from mine waters using
bacterial sulfate reduction. Applied Microbiology and Biotechnology37, 674-678.
Hedin, R.S., Hammack, R., and Hyman, D. (1989). Potential Importance of Sulfate Reduction
Processes in Wetlands Constructed to Treat Mine Drainage. In 'Constructed Wetlands
for Wastewater Treatment- Municipal, Industrial and Agricultural'. (Ed. D.A. Hammer.)
pp. 508-509. (Lewis Publishers: Chelsea, Michigan.)
Hedin, R.S., and Nairn, R.W. (1992). Designing and sizing passive mine drainage treatment
systems. Proceedings Thirteen Annual West Virginia Surface Mine Drainage Task Force
Symposium - April 8-9, 1992.
Hedin, R.S., Nairn, R.W., and Kleinman, R.L.P. (1994) Passive treatment of coal mine drainage.
US Bureau of Mines Information Circular IC-9389, Pittsburgh, PA, USA, 35p.
Hie\, M.T., and Kerins, F.J. Jr. (1988). The Tracey wetlands: a case study of two passive mine
drainage treatment systems in Montana. In 'Mine drainage and surface mine reclamation
(conference proceedings). Volume 1: Mine water and mine waste. Pittsburgh, PA, USA,
19-21 Apr 1988'. US Department of the Interior, Bureau of Mines Information Circular
IC9183. pp. 352-358 (Apr 1988).
Jenne, E.A. (1968). Controls on manganese, iron, cobalt, nickel, copper and zinc concentrations
in soils and water: the significant role of hydrous manganese and iron oxides. In 'Trace
lnorganics in Water'. (Ed. R.A. Baker.) Advances in Chemistry Series No. 73,337-387.
(American Chemical Society: Washington, DC.)
Macfie, S.M., and Crowder, A.A. (1987). Soil factors influencing ferric hydroxide plaque formation
on roots of Typha latifolia L. Plant and Soil1 02, 177-184.
Machemer, S.D., Lemke, P.R., Wildeman, T.R., Cohen, R.R., Klusman, R.W., Emerick, J.C.,
and Bates, E.R. (1990). Passive treatment of metals mine drainage through use of a
constructed wetland. In 'Proceedings of the 16th Annual Hazardous Waste Research
Symposium (EP N600/9-90-037 ed.)'. pp. 104-114. (US Environmental Protection Agency:
Cincinnati OH.)
Masters, B.K. (1988). Biological filtration of industrial effluent water, a successful case study:
Third International Mine Water Congress Oct. 1988, Melbourne, Australia. pp. 763-770.
Mcintire, P.E., Edenborn, H.M., and Hammack, R.W. (1990). Incorporation of Bacterial Sulfate
Reduction into Constructed Wetlands for the treatment of Acid and Metal Mine Drainage:
1990 National Symposium on Mining, May 14-18, 1990. pp. 207.
Morea, S.C., Olsen, R.L., and Chappel, R.W. (1989). Assessment of a passive treatment system
for acid mine drainage at a Colorado Supertund site, Camp Dresser & McKee: Denver.
Nairn, R.W., Hedin, R.S., and Watzlaf, G.R. (1991 ). A Preliminary Review of the Use of Anoxic
Limestone Drains in the Passive Treatment of Acid Mine Drainage. In 'Proceedings
West Virginia Sur! ace Mine Drainage Task Force Symposium (Twelfth Annual ed.)'. (West
JONES & CHAPMAN 144
Wetlands to Treat AMD.
Virginia surtace Mine Drainage Task Force/West Virginia Mining & Reclamation Association:
Morgantown, West Virginia.)
Nairn, R.W., Hedin, R.S., and Watzlaf, G.R. (1992). Generation of Alkalinity in an Anoxic Umestone
Drain. In 'Proceedings of the National Meeting of the American Society for Surtace Mining
and Reclamation, Duluth, Minnesota, American Society for Surface Mining and Reclamation,
ASSMR9-92 June 14-18, 1992'.
Noller, B.N., Woods, P.H., and Ross, B.T. (1992). Case studies of wetland filtration of mine
water in constructed and naturally occurring systems in northern Australia. In 'Proc.
Inti. Specialist Conference Wetlands Downunder-Wetland Systems in Water Pollution
Control, Uni. of NSW, Sydney, NSW, Australia, Nov. 1992'. pp. 47.1-47.9
Patterson, J.W. ( 1985). 'Industrial Wastewater Treatment Technology (2nd ed.).' (Butterworth
Publishers: Stoneham, MA.)
Postgate, J.R. (1979). 'The Sulphur-Reducing Bacteria.' (Cambridge University Press: New
York.)
Ryan, P., and Hosking, R. (1992). A review of wetland filter systems for mining operations.
In 'Proceedings of the International Specialist Conference on Wetlands Down under -Wetland
Systems in Water Pollution Control, Uni. of NSW, Sydney, NSW, Australia, Nov. 1992'.
pp. 1-16.
Reed, S.C., Middlebrooks, E.J., and Crites, R.W (1988). 'Natural Systems for Waste Management
and Control'. (McGraw Hill Book Company: Sydney.)
Sencindiver, J.C., and Bhumbla, O.K. (1988). Effects of cattails (Typha) on metal removal from
mine drainage. Bureau of Mines Information Circular, 9183, pp. 359-366.
Skousen, J.G., Phipps, T.T., and Fletcher, J. (1992). Acid mine drainage treatment alternatives.
In 'Land reclamation: advances in research and technology- proceedings of the international
symposium, Nashville, TN, USA'. (Eds T. Younos, P. Diplas, and S. Mostaghimi.) pp.
297-303. (American Society of Agricultural Engineers: St Joseph, Ml USA.)
TarutisJr, W.J., and Unz, R.F. (1990). Chemical diagenesis of iron and manganese in constructed
wetlands receiving acidic mine drainage. In 'Constructed Wetlands in Water Pollution
Control'. (Eds. P.E. Cooper and B.C. Findlater.) pp. 429-440. (Permagon Press: Oxford.)
Wildeman, T.R., and Laudon, L.S. (1989). Use of Wetlands for Treatment of Environmental
Problems in Mining: Non-Coal-Mining Applications. In 'Constructed Wetlands for
Wastewater Treatment- Municipal, Industrial and Agricultural'. (Ed. D.A. Hammer.) pp.
221-231. (Lewis Publishers: Chelsea, Michigan.)
Wildeman, T., Brodie, G., and Gusek, J. (1992). Wetland Design for Mining Operations, Colorado
School of Mines.
Wieder, R.K. (1989). A survey of constructed wetlands for acid coal mine drainage treatment
in the eastern United States. Wetlands 9, 229.
145
JONES & CHAPMAN
146
Second Auslralian Acid Mine Drainage Workshop (Eds N.J. Grundon and L.C. Bell), pp. 147-164.
ASSESSMENT OF PROTECTIVE COVERS AND DUMP BEHAVIOUR
G.W. Wilson
Unsaturated Soils Research Group, Department of Civil Engineering,
University of Saskatchewan, 57 Campus Drive, Saskatoon,
Saskatchewan, Canada, S?N 5A9
ABSTRACT
A research program involving field instrumentation, laboratory testing and numerical modelling
for the evaluation of soil covers installed on waste dumps, at mine sites in Canada and the United
States, has been completed. One site is located in a humid environment, whereas, the second
site is situated in an arid climate. The objective of the research was to evaluate the performance
of soil covers with respect to water and oxygen fluxes. A newly developed soil/atmosphere
flux boundary model was used to predict mass transport rates. The analysis shows that the
cover installed at the humid site functions as an oxygen and infiltration barrier. The cover for
the arid site reduces infiltration to values approaching zero.
Excavation of a large section of the waste dump (15 Mt) at the arid site has also been carried
out. The dump was found to be well structured with respect to particle size distribution. This
structure appears to strongly influence the heat and mass transport processes within the dump.
1.0 INTRODUCTION
Protective covers for waste rock are generally the last design option which can be implemented
to reduce Acid Rock Drainage (ARD). These covers form the uppermost control surface or
boundary which delivers water and oxygen to the underlying waste rock. The use of protective
covers for waste rock dumps is gaining popularity in Canada and the United States. In gen(jral,
however, analytical techniques and design methodologies are still developing. Performance-based
experience is limited and the effectiveness of these cover systems in preventing or reducing
ARD can be questioned. Some of the issues which must be considered for the design of a
protective cover system are climate and surface hydrology, hydraulic and thermal properties
of the soil materials, hydraulic characteristics of the waste rock and groundwater flow conditions,
vegetation, slope and erosional integrity, and long term ecological stability.
This paper provides an assessment of the protective covers and dump behaviour for two waste
rock sites. Site I is located in the interior of the province of British Columbia, Canada, while
Site II is located in the state of Montana, United States. A research program to assess the
performance of the protective covers installed at these sites is currently in progress. The research
program involves the installation and monitoring of field instrumentation, the measurement of
the hydraulic and thermal properties of the cover materials and waste rock, numerical simulation
of field measurements, and long-term predictive modelling. This paper summarises the major
findings of the research completed to date and provides an assessment of performance. In
addition, early results of a second research program involving a large excavation of the waste
rock dump at Site II is included as it relates directly to dump behaviour.
147 WILSON
Protective Covers and Dump Behaviour.
2.0 BACKGROUND
A newly developed computer program called SoiiCover was developed at the University of
Saskatchewan as a computer-aided design tool for evaluating the performance of protective
soil covers. This work was carried out under the Canadian MEND (Mine Environment Neutral
Drainage) program. Placer Dome Inc. subsequently extended the development and application
of this software. The model is used here to evaluate the performance of the recently constructed
soil cover systems at the two waste rock dumps. The location of these sites is shown in Fig. 1
as Sites I and II.
Equity Mine (Site I) Golden Sunlight Mine (Site II)
Fig. 1. Location of Site I and Site II.
The selection of the sites was not arbitrary. These sites were chosen on the basis of climatic
conditions. Site I is considered to represent a humid environment (i.e. precipitation exceeding
evaporation), while Site II is located in an arid environment (i.e. evaporaUon exceeding precipitation).
These two environments impose different performance criteria for the soil cover systems. In
general, all soil covers may be designed as infiltration barriers to reduce drainage through the
waste rock. However, it may also be possible to design covers to function as oxygen barriers
(Yanful et al. 1993) which will reduce oxidation of the sulfide minerals.
The protective cover installed in the humid environment of Site I was designed principally as
an infiltration barrier. The designers also believed, however, that it may be possible for the
cover to maintain a high degree of saturation overtime due to the high precipitation and relatively
low evaporative demand. This would permit the cover to function as an oxygen barrier. The
cover installed at Site II was also designed as an infiltration barrier. However, due to the aridity
of the site, performance as an oxygen barrier was not considered to be a potential benefit.
Therefore, optimal performance with respect to minimisation of net infiltration was considered
paramount. The performance of the two cover systems with respect to net filtration, moisture
saturation and oxygen diffusion are the primary issues which will be discussed in this paper.
2.1 Site I
Site I is located in the semi-humid alpine environment of central British Columbia. The mean
annual precipitation is approximately 710 mm, 43% of which occurs as rainfall while 57% occurs
as snow. The total annual potential evaporation is estimated to vary between 300 mm and 500
mm per year which is well below the total precipitation. The average monthly temperature for
WILSON
148
Protective Covers and Dump Behaviour.
the site is below zero for the months of November through March. Snowpack accumulation
can exceed 350 mm (water equivalent), the majority of which melts during the month of May.
A maximum average temperature of 20.3 oc occurs during the month of June. Average wind
speeds vary little between 6 km h-' and 8 km h-' throughout the year, while average monthly
relative humidities range between 43% and 80%.
Site I has three dumps located adjacent to each other which contain approximately 85 Mt of
waste rock. The dumps were initially constructed by end dumping and later in lifts of approximately
10 m. The waste rock is estimated to contain between 2% and 3% pyritic sulfide. Acidic drainage
is currently being collected and treated with lime. Construction of a 0.8 m thick till cover began
in 1990 in progressive stages with completion in late 1994. Seeding with clover and grasses
has been completed and vegetative stands are developing well.
2.2 Site II
Site II is located within an arid environment in the Northwest United States. The mean annual
precipitation is approximately 240 mm year' with about 80% occurring as rainfall and the remainder
as snow. Potential evaporation is estimated to be approximately 750 mm year' on the basis
of measured pan evaporation. This value greatly exceeds precipitation. Mean monthly temperatures
are below zero for the months of December through March with the highest average maximum
of 29 oc occurring in July. Mean monthly wind speeds vary between 1 o km h-' and 15 km h-',
and the mean monthly relative humidity for the summer months is approximately 50%.
Approximately 200 Mt of waste rock has been deposited in two large dumps over the past 12
years. The predominantly shale waste rock contains disseminated and massive sulfides which
are highly reactive and as a result, oxidisation occurs rapidly. The largest dump is approximately
100 m in height, and was constructed by end dumping. Oxidisation is known to be occurring
within the dump where temperatures in excess of 65 oc have been measured. Numerous vents
have been observed at locations throughout the dump. However, acidic drainage from the dump
has not been observed, and collection and treatment has not been required to date. Reclamation
of the completed areas of the dumps is under way. Regrading of the side slopes to a reduced
slope of 2:1 was completed prior to the construction of a 1.2 m thick layered protective cover.
The cover system includes 0.6 m of topsoil overlying 0.6 m of oxide cap material. The oxide
cap is essentially fully weathered waste rock and the topsoil layer is made up of natural surface
soils which have been excavated and stockpiled for reclamation prior to dump placement. A
number of test plots for the cover system were previously installed and instrumented. Seeding
with native grasses was successful and vegetation stands are established on the test plot areas.
3.0 THEORY AND NUMERICAL MODELLING
The numerical modelling of water, heat and oxygen fluxes through the cover systems was carried
out using software based on Soil Cover Version 1.0 (MEND, 1993). The computer program solves
for one-dimensional, transient, coupled heat and mass transfer through soil cover systems based
on atmospheric forcing conditions and hydraulic head and temperature conditions at depth.
Wilson et al. (1994) provides the equations for coupled heat and mass transport. Mass transfer
includes liquid water, water vapour and oxygen. The equation for the flow of liquid water and
water vapour is given as follows:
ohw ~ C ~ ~ (kw OhwJ+ c;_, ~ (Dv oPvJ [1]
ot oy ~ oy oy ~ oy
where hw is total head, tis time, C ~ is the modulus of volume change with respect to the liquid
149
WILSON
Protective Covers and Dump Behaviour.
water phase, y is position, kw is hydraulic conductivity, C0 is the modulus of volume change
with respect to the water vapour phase, Dv is the diffusion coefficient of water vapour through
the soil, and Pv is the actual vapour pressure within the unsaturated soil voids. The vapour
pressure within the soil is calculated on the basis of total suction, ljf, in the liquid phase as:
[2]

g w,J
P-efiTP
v sv
where g is acceleration due to gravity, Wv is the molecular weight of water, R is the universal
gas constant, Tis temperature, and Psv is the saturation vapour pressure at temperature T.
Heat flow is coupled with the mass transfer given by Eqn. [1] and is computed as:
C OT _ (1.. 8TJ- L ( P + (
0
8Pvl
h 81 oy ( 8y v ( P By v By)
[3]
where Chis the volumetric specific heat of the soil as a function of water content, 1.. is the thermal
conductivity of the soil as a function of water content, Lv is the latent heat of vaporisation, and
pis the total gas pressure in the air phase of the soil. The finite element solution for the coupled
system of equations is given by Joshi et al. (1993).
The upper boundary conditions for the heat and mass transfer equations are defined through
coupling with climatic conditions. Precipitation events are specified as a positive flux boundary
condition. Should this flux exceed infiltration capacity, the excess is set equal to runoff. Actual
evaporation depends on both atmospheric forcing conditions and the actual vapour pressure
at the surface of the soil cover and is determined using the modified Penman formulation proposed
by Wilson (1990) as follows:
E- 1'.0 + y.
6.+A
[4]
where E is evaporative flux, 1'. is the slope of the saturation vapour pressure versus temperature
curve at the temperature of the air, Q is net radiation, y is the psychrometric constant, and E.
is equal to f(u) e.(B- A), where f(u) is a function dependent on wind speed, surface roughness
and eddy diffusion, e. is the vapour pressure in the air above the soil surface, B is the inverse
of the relative humidity of the air, and A is the inverse of the relative humidity at the soil surface.
Temperature at the surface of the soil, T sis simultaneously defined using the relationship proposed
by Wilson (1990):
T
8
T. + (_
1
-J(O- E)
(r
where T. is the temperature of the air above the soil surface.
[5]
Equation [ 4] for computing the rate of evaporation is used for the special case of a non-vegetated
soil cover. Actual evapotranspiration is calculated by including a transpiration flux computed
using a modified form of the method proposed by Ritchie (1972). Tratch (1995) modified the
function given by Ritchie (1972) such that the potential transpiration is reduced by a plant limiting
factor if the values of matric suction in the root zone exceed 100 kPa.
In summary, the system of coupled equations listed are used in SoiiCover to compute heat and
moisture fluxes through the soil cover system on the basis of climatic parameters including
precipitation, all net radiation, air temperature, relative humidity, wind speed and vegetation
factors such as leaf area index, root density and root depth. Soil, vegetation and atmospheric
WILSON
\50
Protective Covers and Dump Behaviour.
conditions are integrated to form a coupled soil/atmosphere continuum such that the surface
flux boundary condition for evapotranspiration depends on both climatic demand and soil suction.
The oxygen flux through the soil cover is determined on the basis of the degree of saturation
in the soil cover profile. The flux of oxygen is computed assuming that steady state conditions
exist for each day. Assuming an oxygen concentration of 21% at the surface of the cover and
zero at the base, the mass flux of oxygen, J, is determined on the basis of Fick's First Law as
follows:
J- 0 6.C
o 6.x
[6]
where Dais the coefficient of oxygen diffusion as a function of the degree of saturation, /'.C is
the change in oxygen concentration, and l'.x is the change in elevation. The coefficient of oxygen
diffusion may be measured experimentally but for this study the value was calculated as a function
of the degree of saturation using the empirical relationships developed by Millington and Shearer
(1971) and Collin and Rasmussen (i 988). The application of SoiiCover to the protective soil
cover systems at Sites I and II as well as the numerical results will be described in the next
sections.
4.0 SITE I
An extensive field instrumentation program for the protective cover system at Site I was initiated
in August 1992 as part of the research program. Instrumentation for the measurement of oxygen
profiles and temperatures within the waste rock dump was previously installed. A total of 7
oxygen probes and 1 0 temperature probes were installed at various locations and depths.
Furthermore, 14 lysimeters were installed at various locations under the protective cover as
construction of the cover proceeded. The lysimeters were installed in the waste rock at the
base of the cover to measure drainage from the cover to the underlying waste. The specific
details for this instrumentation are not included in this paper in the interest of space; however,
details for the instrumentation installed in the cover itself are described below.
4.1 Field Instrumentation and Material Properties
The instrumentation installed in the cover system consists of thermal conductivity sensors for
the in situ measurement of matric suction and temperature, neutron probe access tubes for
the in situ measurement of volumetric water content and a weather station for the measurement
of climatic conditions. Figure 2 shows the location of the main instrumentation installed at Site
I. The key features for the cover instrumentation are the weather station on the Main Dump
and the three culvert stations located on the Main Dump and Southern Tails Dump. The purpose
of this instrumentation was to measure moisture and temperature conditions in the profile of
the protective cover along with the climatic conditions which control heat and mass flow in the
soil profile. The weather station on the Main Dump is fully automated to provide continuous
measurement of precipitation, global and net radiation, air wind speed and relative
humidity. Measurements are recorded each minute and averaged to give mean hourly values.
The weather station contains a data logger which stores all records for a period of up to 3 months
at which time it is down loaded.
The three access culvert stations are installed on the top of the Main Dump (TMD), on the South
West Face (SWF) of the main dump and on the Southern Tails Dump (STD). Figure 3 shows
a cross-section for the culvert and sensor installations. Each culvert measures 0.9 min diameter
and 2m in length. The culverts extend through the protective soil cover and into the underlying
waste rock.
151
WILSON
Protective Covers and Dump Behaviour.
. ~ \
) ) ~
.# / / - ~
,. .../ l/"' \ ... ?
LEGEND
X WS - Weather Station
TMD -Top of Main Dump Site
\
>
'ft. ....
---. (
SWF - Southwest Face of Main Dump Site
STD -Southern Tails Dump Site
0 NP-X - Neutron Probe Access Tube
-- Crest
-- 20 m Contour
-Toe
Fig. 2. Site plant showing the location of field
Instrumentation at Site I (from O'Kane 1995).
Eight thermal conductivity sensors
are installed in the cover and waste
rock materials at a distance of 0.6
m from the edge of the culvert. The
thermal conductivity sensors consist
of a 2 ern diameter porous ceramic
stone which is placed in direct
contact with the soil cover material.
The water content in the ceramic
comes to equilibrium with the pore
water pressure in the soil. When
the pore water pressure is negative,
which is almost always the case, the
water content of the stone will be
at a value of less than 100%
saturation. A small heating element
within the stone is used to generate
a heat pulse and the decay rate is
measured by a thermocouple. The
rate of cooling provides a measure
of thermal conductivity, water content
and matric suction. A calibration
curve for each sensor is determined
which allows for the measurement
and calculation of the matric suction
in the soil at the tip of the porous
stone. The thermal conductivity
sensors are connected to a data
acquisition system and continuous
readings of matric suction and soil
temperature are obtained and stored
for subsequent retrieval.
In addition to the three culvert
systems installed at Site I, 14
neutron access tubes were also
installed at various locations to
provide greater coverage for
moisture conditions of the soil cover
through the total area of the three
waste dumps. These access tubes
were pushed through the cover layer to the underlying waste rock. A neutron source was manually
inserted down the tubes to give a continuous profile of water content. A calibration curve for
each access tube was produced by retrieving small samples of the soil materials adjacent to
the walls of the tubes for the direct determination of water content. Moisture profiles were obtained
for the access tubes at various times and compared to the measurements obtained at the culvert
station in order to assess spatial performance of the cover. Some of the measurements for
matric suction, water content and temperature obtained from the culvert stations will be presented
in a subsequent section.
WILSON
152
DATA ACQUISITION
AND
SOLAR PANEL
THERMAL CONDUCTIVITY
I
I
RS
NOT TO SCALE
Fig. 3. Cross-section of the access culverts
and sensor Installations for Site I (from
O'Kane 1995).
0.26
~ 0.22
C'
8 0.18
l!l
~ 0.14
~
0.10 'l!)
.!;;
~ 0.06
0
0.02
1e2 1e4 1e6
Suction (kPa)
Fig. 4. Soil water characteristic curves for the
uncompacted till, compacted till, and
waste rock (after Swanson 1994).
153
Protective Covers and Dump Behaviour.
The till material used to construct the
protective cover at Site I is classified as
SC-CL according to the Unified Soils
Classification System with approximately
23% cobble and gravel size, 28% sand,
40% silt and 9% clay size particles. The
till for the lower 0.5 m of the cover was
compacted to a minimum of 95% of the
saturated proctor density of 1.9 t m_, at
an optimum gravimeter water content of
approximately 14%. The same till material
was used for the upper 0.3 m of the cover;
however, this upper section was not
compacted. This effectively created a
layered cover system. The soil water
characteristic curves, SWCC, for the
compacted and non compacted tills are
shown in Fig. 4. The effect of compaction
on the SWCC relationship can be seen
in Fig. 4 with respect to an increase in
the air entry value, AEV (i.e. approximately
1 00 kPa versus 1 0 kPa), and the reduction
in the saturated water content (i.e. 0.20
versus 0.22) for the compacted till. The
SWCC used to represent the waste rock
underlying the cover is also shown in
Fig. 4. Figure 5 shows the relationship
between hydraulic conductivity and matric
suction computed for each of the materials
based on the SWCC curves in Fig. 4 using
the commercially available software
package called KCAL (Gee-Slope Int.
1993). Relationships between thermal
conductivity and specific heat capacity
versus water content were also
determined for the uncompacted till,
compacted till and waste rock using the
commercially available computer program,
TheHyProS (Tarnawski and Wagner,
1992).
The hydraulic and thermal properties
described above were used in the
computer program SoiiCover to simulated
flow through the layered system of
uncompacted till, compacted till and waste
rock. Climatic parameters obtained from
the weather station were used as input
parameters. The output from the model
simulations was then compared to
measured values obtained from the field
instrumentation.
WILSON
Protective Covers and Dump Behaviour.
1e-04
::'
E
.g_ 1e-08
.
.2:
u
]
1e-12
.Q
1e-16 ""5

"0
f
te-20
1e0 1e2 1e4 1e6
Suction (kPa)
Fig. 5. Unsaturated hydraulic conductivity
versus matrlc suction for the cover
materials and waste rock (after Swanson
1994).
4.2 Modelling and Assessment of
Cover Performance
The TMD culvert site was selected for
the model simulations because it was
located next to the weather station_ This
area is gently sloping to flat and was
sparsely vegetated at the time. A
simulation period of approximately 150
days commencing at the beginning of
June 1993 was selected. Initial water
contents and temperature profiles through
the cover used for the transient modelling
were obtained from the culvert station
and a nearby neutron access tube.
Figure 6 shows the surface flux boundary
condition for the simulation period. The
positive fluxes represent precipitation
events measured at the weather station,
while negative fluxes are evaporation
events intermittent to periods of rainfall. The values of evaporation were computed by SoiiCover
using the Modified Penman formulations given in Eqn. [4].
Figure 6 also shows the cumulative water balance for the simulation period. The total precipitation
for the period was approximately 360 mm with runoff computed to be slightly less than 100 mm.
The actual evaporation was computed to be 290 mm which exceeded the total precipitation
less runoff. Hence, a decrease in soil moisture storage within the cover profile should have
occurred.
It should be noted that the cumulative value of AE is less than the cumulative value of PE which
exceeds 400 mm. During the initial period of June, AE was approximately equal to PE since
precipitation was adequate to keep soil moisture conditions wet. The cumulative AE fell below
the PE as the simulation continued since precipitation was not sufficient to maintain the wet
soil conditions and high evaporation. Had the values of PE been used to define the upper flux
boundary condition in the simulation, excessive soil moisture depletion would have been calculated.
"E 20
E
; 15
::J
u:: 10
(!)
5 0
rei
'I::
0
::J
(J)
- -5
CD
Z-10
June July Aug. Sept Oct.
"E 4oo
..
2.l 200
c
rei
(ij 0
m
li5 -200

-400
E
8 June July Aug. Sept. Oct.
Fig. 6. Surface flux boundary condition for Site I (after Swanson 1994).
WILSON 154
Protective Covers and Dump Behaviour.
Figure 7 shows the computed and measured values of matric suction for a depth of 13 em and
18 em in the loose till. These shallow depths were selected for comparison because they show
the most dynamic response in suction as compared to those for the lower compacted till. It
can be seen from Fig. 7 that there is a good correlation between measured and computed values
for matric suction. The response is most significant during the drying period in September and
October where suctions increase to high values.
ro-

c
.2
0
200
()

0
.
-
1 Computed Matric Suction 1
o Measured Matric Suction
I
I 1/ /
/l v
18 em I
H 13 em I
I
/;
A
.I l_y.,

/.....-: .-/

_,a ...,.
ln. II

_,.
May June July Aug. Sept. Oct. May June July Aug. Sept Oct.
Flg. 7. Measured and computed values for matrlc suction versus time at two depths In
the soli cover at Site I (after Swanson 1994).
Figure 8 shows the computed and measured values of soil temperature at depths of 13 em and
31 em in the soil profile. Good agreement between the measured and computed temperatures
is apparent.
20


::J
!
& 10
E

c;
(J)
0
a,
'b
May June July Aug. Sept. Oct. May June July Aug. Sept. Oct.
Fig. 8, Measured and computed values of soil temperature within the cover system at
Site I (after Swanson 1994).
Figure 9 illustrates the effect of vegetation on the water content profile in the protective cover.
The measured and computed water contents correspond to bare and vegetated soil surfaces
at neutron access tube locations NP 1 and NP 2 near the SWF culvert station (Fig. 2). The
results plotted are for the month of September which was the most arid. It can be seen that
the water content profile corresponding to the vegetated site is the driest as one would expect.
However, it is interesting to note that the water content profile in the lower compacted till layer
remained high at approximately the saturated water content of 20%. The removal of water from
the protective cover occurred in the upper loose layer leaving the lower compacted layer saturated.
This is a positive result since the maintenance of saturated conditions in the lower section of
the cover will help minimise the diffusion of oxygen to the underlying waste rock.
155 WILSON
Protective Covers and Dump Behaviour.
1.8
1.6
.s
c
0
"iii 1.4
>
.Jl1
w
1.2
-------
\..
.-:--


y
I I



o Sept. 30
I<> Aug. 121
r\\i' .
t. Aug. 31
..
j - Computed I
j.
[ Measured 1
J R Vegetated Cover I

H Bare Soil
I
:
:
1
0 5 10 15 20 25 0 5 10 15 20 25
Gravimetric Water Content % Gravimetric Water Content %
Fig. 9. Measured and computed water content profiles for bare soli and areas with protective
vegetative cover at Site I (after Swanson 1994).
'"'' 1e-1
E
.s
c
CD
;g 1e-3
CD
0
0
c
0
'gj 1 e-5
:1::
0
c
CD
Ol
S( 1e-7
0 0
Fig. 10.
0.2 0.4 0.6 0.8 1.0
Degree of Saturation
Oxygen diffusion coetflclent
versus degree of saturation
for till cover at Site I (after
Swanson 1994}.
Estimates of oxygen flux through the protective
cover at Site I were conducted with SoiiCover
based on the computed water content profiles
and Eqn. [6]. Figure 1 0 shows the relationship
between the oxygen diffusion coefficient and
the degree of saturation used for the analysis
determined on the basis of experimental data
and the Millington and Shearer (1971) method.
The oxygen concentration was assumed to
be atmospheric at the surface of the cover
and zero at the base of the cover or top of
the waste rock. Simulations were carried out
for dry, mean and wet years and were
determined on the basis of regional climate
data extending over a fifty year period from
1943. The dry, mean and wet years
corresponded to 343 mm, 549 mm and
760 mm of total annual precipitation. The flux
of water infiltrating through the cover was also
computed for the same periods. The mass
flux calculations for Site I showed that the
amount of water infiltrating through the cover
to the waste rock was < 3 mm year' for all three typical years. In other words, less than 1
%of all precipitation entered the waste rock dump. This compares to 203 mm (60%). 371 mm
(68%), and 593 mm (78%) for the dry, mean and wet years, respectively, for bare or uncovered
waste rock. Infiltration through the cover is controlled largely by the low saturated hydraulic
conductivity of 2 x 1 o-" em s' for the compacted till. Increasing this value by one order of
magnitude increased the infiltration somewhat, but in general did not increase it to values exceeding
5% of the total annual precipitation. This maximum value corresponded to the bare soil condition.
Oxygen fluxes were computed to be low for each of the dry, mean and wet years. The maximum
flux of oxygen through the protective cover occurred in the dry year and with vegetation. This
value was computed to be 0.02 kg m-' year-'. Increasing hydraulic conductivity by 1 0-fold resulted
WILSON 156
Protective Covers and Dump Behaviour.
in a higher flux of 0.148 kg m' year'. These numbers are considerably lower than those computed
for the diffusion of oxygen into the uncovered waste rock (range 7.695 kg m' year' to 9.232
kg m' year-'). In summary, the analysis shows the oxygen fluxes to be reduced by a factor
of at least 50 and more typically in the order of 400 when the protective cover is installed. The
analyses described above indicate that the installation of the protective cover at Site I will
significantly reduce the flux of water and oxygen entering the wasle dump. It is difficult to fully
assess performance at this time since full closure with the cover has only recently been completed.
However, early trends indicate that lime quantities for the treatment of the acidic drainage have
decreased by up to 30%. Furthermore, total discharge quantities have also decreased significantly.
Fig. 11.
c
2
c
0.4
l5 0.3
Q;
0.2
.!2
Instrumentation layout
(neutron probes and erosion
troughs only) for the High
Grade Waste test plots (after
Swanson 1994).
-
"''
L Topsoil Cover}


j Waste Rock
/f
0.1
E
- Oxide Cap 1


J :::J
g
0
1e-2
Fig. 12.

1e0 1e2 1e4 1e6
Suction (kPa)
Soil water characteristic curves
for the topsoil, oxide cap, and
waste rock materials used at Site
II (after Swanson 1994}.
157
5.0 SITE II
The field instrumentation installed at Site II
was installed by the mine operator prior to the
initiation of work for the research program
described here. The instrumentation used
for this study was principally the neutron
access tube for the measurement of in situ
water content through the cover. The neutron
access tubes were installed on the level dump
top and on the test slopes which were
regraded to 2:1 and 3:1 slopes as shown in
Fig. 11. Erosion control is a potential concern
and erosion troughs were installed to monitor
soil losses. A weather station similar to that
described for Site I was also installed at a
location near the test plots.
5.1 Material Properties and Numerical
Modelling
The protective cover installed at the test
area consists of 0.6 m of topsoil over 0.6 m
of an oxide cap. The oxide cap is
essentially fully weathered waste rock with
little or no acid generating potential. The
top soil layer, as previously stated, is
comprised ol natural soil horizons which
were excavated prior to dump placement
and stockpiled for final reclamation.
Figure 12 shows the SWCC for the topsoil,
oxide cap and undertying waste rock. The
SWCC relationships for the three materials
are quite similar. The waste rock, oxide
cap and topsoil all have a relatively low
AEV ranging between approximately 1 kPa
and 3 kPa. The most significant variation
in the SWCC relationships is the slope of
the curves at values of matric suction
exceeding the AEV. The SWCC
relationship for the waste rock drops most
WILSON
Protective Covers and Dump Behaviour.
1e-1

D Topsoil
E
!:!.-
1e-5


<> Oxide Cap
.A. Waste Ro
.c
:;;;
= u
:::>
1e-9
u
c:
0
()
.!2
1e-13
u
:t
1e-17
1e-1
Flg.13.


'
I
'

=1
"'
i
'
I
I
"
I
- --- -- ---- ---
1e1 1e3 1 e5
Matric Suction (kPa)
1e7
Unsaturated hydraulic
conductivity functions for the
topsoil, oxide cap, and waste
rock used at Site II (after
Swanson 1994).
rapidly when compared to the oxide cap
and topsoil. This suggests that the oxide
cap and topsoil have a more uniform
gradation. In general, the SWCC
relationships for the topsoil and oxide cap
appear to be a slightly modified form of
the curve for the waste rock. This is a
logical result since all materials are
basically derivatives of the same parent
rock.
Figure 13 shows the hydraulic conductivity
versus matric suction relationship for each
material calculated on the basis of the
SWCC relationships shown in Fig. 12.
These properties along with the
relationships determined tor thermal
conductivity and heat capacity versus water
content determined using TheHyProS
(Tamawski and Wagner, 1992) were used
for numerical simulations in Soi!Cover.
The section of the test cover simulated
was the level top of the dump at Neutron
Probe NP-1 (Fig. 11). This area was
vegetated, hence, transpiration fluxes were included in the analysis.
Figure 14 shows the computed surface fluxes used for the period of May 1 through mid November
of 1992. Total precipitation, most of which occurred before mid July, was approximately 225 mm.
Runoff is shown to be insignificant while actuaJ evapotranspiration is slightly higher than precipitation
and equal to approximately 230 mm. Potential evaporation for the period greatly exceeds actual
evapotranspiration (i.e. 820 mm versus 230 mm}.
>: 30
25
i 20
15
u:: 10
(])
1il 5
't:
J5 0
Q) -5
z
-10

I I .1 I I
r r
rr
-
L
1.JL
]
I I I
200 =
(])
u
c
(1j
(ij
Cil
"-
(])
-400
1 --+I -+! ...::-,m*------+PE----1--l
May June July Aug. Sept. Oct.
0
May June July Aug. Sept. Oct.
Flg.14.
WILSON
Calculated dally net surface flux and cumulative water balance components
for Site II In 1992 (after Swanson 1994).
158
'
Protective Covers and Dump Behaviour.
Figure 15 shows the measured and computed water content profiles through the topsoil, oxide
cap and waste rock for the same period shown in Fig. 14. Reasonably good agreement can
be seen between the measured and computed values of water content. In both cases, the oxide
cap shows lower water contents than the topsoil above and waste rock below. This can be
attributed to the slight differences in the SWCC for each material which creates a somewhat
layered system. Little change in the water content profile occurred over the simulation period.
In general, the profiles show that progressive drying occurred during the period which is consistent
with the surface fluxes shown in Fig. 14.
Figure 15 also shows the water content profile for the test cover system to be unsaturated for
the entire period of the simulation. Comparing measured and computed water contents for each
material with the SWCC relationships shown in Fig. 12 indicates degrees of saturation ranging
between 50% and 60%. It is clear, based on this observation, that the protective cover will not
function as an effective oxygen barrier. The primary focus tor the subsequent analysis is therefore
directed towards the evaluation of the net infiltrative fluxes which will enter the waste rock through
the cover system.
3.5
3
I 2.5
c:
0
-.:;

2
(])
iii
1.5
0
Fig. 15.
-o- 8/6/92
--9/17/92
5 10 15 20 25 0 5 10 15 20 25
Volumetric Water Content Volumetric Water Content
Measured and computed water content profiles for Site II (from Swanson
1994).
Predictive modelling cases to estimate net infiltrative fluxes were carried out for the mean year
and the extreme wet and dry years with respect to total precipitation. Regional climatic records
for a period of approximately 75 years extending back to 1920 were used. The extreme dry
year (1924} had a recorded precipitation of 159 mm, while the extreme wet year (1993), also
the most recent year on record, had a total precipitation of 509 mm. The mean year (1973)
precipitation was determined to be 263 mm.
In addition to varying climatic conditions, the effect of varying the quality of vegetation and hydraulic
conductivity of the cover layers was also evaluated. Table I summarises the results of the analysis.
The base case used the material properties previously described. The good vegetation case
was the same as the base case except the root depth was extended to 80 em from 60 em and
the leaf area indices (LA!) for a poor grass stand were increased by a factor of two to represent
159 WILSON
Protective Covers and Dump Behaviour.
a good grass stand (Schroeder et al. 1984). Simulations were also performed with the topsoil
and oxide cap layers hydraulic conductivity values increased and decreased by one order of
magnitude compared to the values used for the base case as shown in Fig. 13.
Table 1. Annual Infiltration rates to the waste rock for Site II.
Scenario
Amount (mm year-') Percent of annual precipitation
Dry year Mean year Wet year Dry Mean year Wet
159 mm 263 mm 509 mm year year
Covered: Base case Upward 0 50 0 0 10
net flux
Covered: Good vegetation Upward Upward 35 0 0 7
net flux net flux
Cover layers: High k Upward Upward 0 0 0 0
net flux net flux
Cover layers: Low k 1 3 30 6
Uncovered waste rock 12 20 80 7 7 16
Final simulations were conducted for the case of uncovered waste rock. The SWCC relationship
for the waste rock shown in Fig. 12 was used; however, the values of saturated hydraulic
conductivity shown in Fig. 13 were increased from 1 x 1 0_. em s-' to 1 x 1 o-' em s-' for suctions
less than 4 kPa. The SWCC relationship shown in Fig. 12 for the waste rock represents a relatively
fine-grained material; however, field studies have shown the waste rock to also contain coarse
materials which daylight at the surface. The hydraulic conductivity for the waste rock was therefore
increased by three orders of magnitude for the simulation in an attempt to emulate the higher
bypass permeability associated with the coarse-grain macro-pores.
Table 1 shows that the maximum infiltration into the waste rock occurred for the uncovered case
with 12 mm, 20 mm and 80 mm of infiltration entering the waste rock for the dry, mean and
wet years, respectively. These values correspond to between 7% and 16% of total precipitation.
The quantity of infiltration through the cover to the waste rock was found to be extremely small
and even negative for the cases with dry and mean year precipitation. Infiltration through the
covered systems was produced only for the wet year. Approximately 50 mm or 10% of the
total precipitation infiltrated through the base case cover. Improving the quality or density of
vegetation reduced this quantity to 35 mm. In addition, decreasing the hydraulic conductivity
of the cover layers decreased infiltration to 30 mm or 6% of the total wet year precipitation.
It is interesting to note, however, that the simulations showed that increasing the hydraulic
conductivity by one order of magnitude decreased the infiltration to zero. At first glance, this
appears erroneous. A higher hydraulic conductivity should allow infiltration to penetrate more
rapidly. However, the increase in hydraulic conductivity also enhances upward flow or exfiltration
due to evapotranspiration. Since the evaporation events extended over prolonged periods, the
increased evaporation capacity of the cover layers more than offset the increase in infiltration
rates which occurred during short precipitation events. Hence, overall net infiltration was reduced.
In summary, the numerical simulations suggest that very little infiltration will enter the waste
rock with the protective cover installed. In addition, only a moderate amount of infiltration will
enter the waste rock in an uncovered state. The mean net infiltration into the bare or open waste
rock has been computed to be 20 mm per year. This represents the average expected infiltration
into the waste rock prior to reclamation.
WILSON 160
Plate 1. Exposed benches of waste rock, Site II.
Plate 2.
Close-up of exposed waste rock, Site II.
Plate 3. Newly deposited waste rock at Site II.
161
Protective Covers and Dump Behaviour.
5.2 Waste Rock Excavation
and Dump Assessment
Over the past several months, a
large section of the East Dump
at Site I was excavated and
moved to another location. Shear
stresses induced by the surcharge
load of the dump caused the
reactivation of a deep seated
landslide block measuring
approximately 300 hectares in
area. Approximately 15 Mt of
waste rock was removed from the
crest of the landslide area to
unload and stabilise the landslide.
The excavation of a vertical
section of the dump with a total
height of approximately 1 00 m
provided an excellent opportunity
to investigate the behaviour and
oxidation processes within the
dump.
Plate 1 shows a cross-section of
the excavated face of the waste
dump. The first three benches,
each approximately 20 m in
height, are shown. The plate
shows the profile of each bench
to be well structured. Beds
dipping at approximately 45are
apparent both in terms of colour
and gradation. This natural angle
of repose occurred during
placement by end-dumping. The
waste rock consists of
sedimentary and intrusive rocks
with disseminated pyrite and
various amounts of massive
pyrite. Oxidation rates vary with
rock type and time.
Plate 2 shows a close-up of the
oxidised waste rock which further
illustrates the coarse and fine
layering. The coarse layers are
mainly slow weathering latite,
while the fine-textured soils are
highly oxidised shale. In contrast,
Plate 3 shows the condition of
fresh waste rock at the time of
WILSON
Protective Covers and Dump Behaviour.
placement. This material appears to be relatively uniform in terms of appearance and particle
size (i.e. rock and cobble). Examination of the exposed benches indicates that the unoxidised
or fresh waste rock weathers to the condition shown in Plates 1 and 2 within a few years.
Construction of the dump began in 1983 with the oldest sections of the dump located in lifts
below those shown in Plates 1 and 2.
A field research program was initiated in October of 1994 to obtain field measurements and
in situ samples of the waste rock. A total of thirty test pits were excavated. Samples of the
individual layers as well as in situ measurements for water content, temperature, matric suction
and relative humidity were obtained. Testing for the samples returned to the laboratory will
be carried out for grainsize distribution, hydraulic conductivity, SWCC and paste pH. This work
is currently in progress, and the results and analysis will be published at a later date; however,
preliminary observations will be discussed here.
All of the waste rock encountered in the test pits and exposed bench faces was found to be
unsaturated. Coarse layers were fully drained while the upper 20 m of the fine layers in the
waste rock were moist with volumetric water contents ranging between 9% and 23% (degree
of saturation ranging between 30% and 75%). This is in strong contrast to the waste rock
encountered at greater depths which was found to be dry with volumetric water contents ranging
between 0.4% and 5%. Dust control problems were frequently encountered during excavation
due to these lower water contents.
Temperatures up to 65 ac were measured in the upper 20m of the waste rock where the water
contents were high, whereas temperatures in the lower dry regions ranged between 15 ac and
25 ac. Numerous vents were observed at the surface of the exposed waste rock. It is believed
that the vapour from these vents originates in the upper 20 m of the waste rock. Excavation
of the vent areas typically showed that a coarse layer of rock was the predominant pathway
tor vapour migration.
In general, the field observations made during the excavation of the waste rock indicate that
seepage is not occurring at the base of the dump. The lower 80 m of the dump was found to
be dry and inactive. The upper 20 m of the waste rock was moist but remained unsaturated.
This moisture may in part be attributed to infiltration through the uncovered surface of the waste
rock. However, it is reported that freshly blasted waste rock can have a high water content.
In any case, the upper most zone remains unsaturated. The principle vertical pathway for the
infiltrating water is the fine layers of waste rock which have the ability to retain water under negative
pore water pressures. The coarse layers are fully drained and appear to serve only as pathways
tor the transfer of water vapour and oxygen.
The reason for the dry condition of the dump at depth has not yet been fully explained. One
possible explanation may be that the water has been previously evacuated through vents when
in situ temperatures were at elevated levels during active oxidation. Active oxidation is apparent
in the upper moist region where steam vents are clearly a mechanism for the discharge of water
to the atmosphere. Furthermore, water is a reactant in the oxidation of pyrite and is therefore
consumed. Numerical modelling for the operative heat and mass transfer processes and
geochemical reactions active in the dump have not been carried out at this time. Subsequent
modelling will be a valuable tool to help explain field observations.
6.0 SUMMARY AND CONCLUSIONS
Field instrumentation, monitoring and numerical modelling tor protective cover systems installed
over waste rock at two sites has been completed. Covers may be installed to restrict oxygen
WIL.SON
162
Protective Covers and Dump Behaviour.
fluxes and/or water fluxes to the waste rock. Results of the research to date show that a fine-
grained cover material can be successfully used in a humid environment to significantly reduce
both liquid water and oxygen fluxes to the underlying waste rock. The analysis shows that liquid
water fluxes were reduced to between 1% and 5% from approximately 70% with the protective
cover installed at Site I. Furthermore, oxygen fluxes were computed to be reduced by a factor
of up to 400 with the addition of the saturated cover. Designing a cover system to function as
an oxygen barrier at an arid site is not an option. However, the protective cover installed at
Site II appears to be functioning as a barrier to infiltration. The analysis indicates that the protective
cover will prevent infiltration for all but the extreme wet years.
The excavation of the waste rock at Site II shows that acidic drainage is not occurring from the
base of the waste rock dump. Numerical analyses predict relatively low rates of infiltration into
the uncovered waste rock equal to approximately 20 mm per year. The upper 20 m section
of the dump is moist but dry conditions prevail at depth. The heat and mass transfer processes
within the dump are strongly controlled by the interbedded structure of the dump. Numerical
analyses for the transfer processes in the dump must be carried out to help explain field
observations.
7.0 ACKNOWLEDGMENTS
All of the field instrumentation, laboratory testing and numerical modelling work for this research
program was performed by Mr. Greg Herasymuik, Mr. Michael 0' Kane and Mr. Darren Swanson
as part of their research towards a Master of Science Degree. In addition, supervision of their
work was supported by Professors Lee Barbour and Del Fredlund who are co-investigators for
this research project. Ms Brenda Bews helped with editing and preparation of the manuscript.
Financial support for the research program was provided by Placer Dome Inc. and the Natural
Sciences and Engineering Research Council of Canada. Mr. Mike Aziz and Mr. Troy Smith
of Placer Dome gave valuable field support and access to historical data. Mr. Keith Ferguson
initiated the research program and provided technical and logistical support. The completion
of this work would not have been possible without the contributions of these individuals and
agencies and their assistance is gratefully acknowledged.
8.0 REFERENCES
Collin, M., and Rasmussen, A. (1988). Comparison of Gas Diffusivity Models for Unsaturated
Porous Media. Soil Science Society of America Journal 52, 1559-1565.
GEO-SLOPE International Ltd. (1993). SEEP!W User's Manual. Calgary, Alberta, Canada.
Joshi, B., Barbour, S.L., Krause, A.E., and Wilson, G.W. (1993). A Finite Element Model for
the Coupled Flow of Heat and Moisture in Soils Under Atmospheric Forcing. Rnite Elements
in Analysis and Design - The International Journal of Applied Finite Elements and Computer
Aided Engineering 15, 57-68.
MEND (1993). SoiiCover User's Manual for Evaporative Flux Model. University of Saskatchewan,
Saskatoon, Saskatchewan, Canada.
Millington, R.J., and Shearer, R.C. (1971 ). Diffusion in Aggregated Porous Media. Soil Science
111 372-378.
O'Kane, M. (1994). Instrumentation and Monitoring of Engineered Soil Covers for Acid Generating
Mine Waste. M.Sc. Thesis, University of Saskatchewan, Saskatoon, Saskatchewan,
Canada.
Ritchie, J.T. (1972). Model for Predicting Evaporation from a Row Crop with Incomplete Cover.
Water Resources Research 8, 1204-1213.
Schroeder, P.R., Morgan, J.M., Walski, T.M., and Gibson, A. C. (1984). Hydrological Evaluation
of Landfill Performance (HELP) Model, Volume I. User's Guide for Version I, International
163 WILSON
II
Protective Covers and Dump Behavtour.
Ground Water Modelling Center.
Swanson, D.A. (1994 ). Predictive Modelling of Moisture Movement in Engineered Soil Covers
for Acid Generating Mine Waste. M.Sc. Thesis, University of Saskatchewan, Saskatoon,
Saskatchewan, Canada.
Tamawski, V. R., and Wagner, B. (1992). A new Computerized Approach to Estimating the Thermal
Properties of Unfrozen Soils. Canadian Geotechnical Joumal29, 714- 720.
Tratch, D. (1995). Moisture Uptake Within the Root Zone. M.Sc. Thesis, University of
Saskatchewan, Saskatoon, Saskatchewan, Canada.
Wilson, G.W., Fredlund, D.G., and Barbour, S.L. (1994). Coupled Soil-Atmosphere Modelling
for Soil Evaporation. Canadian Geotechnical Journal31, 151-161.
Wilson, G.W. (1990). Soil Evaporative Fluxes for Geotechnical Engineering Problems. Ph.D.
Thesis, University of Saskatchewan, Saskatoon, Saskatchewan, Canada.
Yanful, E.K., Bell, A.V., and Woyshner, M.R. (1993). Design of a composite soil cover for an
experimental waste rock pile near Newcastle, New Brunswick, Canada. Canadian
Geotechnical Journal30, 578-587.
WILSON
164
Second Australian ACid Mine Oramage Workshop (Eds. N J. Grundon and L.C. Bell), pp. 165-177
FINANCIAL AND LONG TERM LIABILITY ASSOCIATED WITH AMD
G.S.C. MurrayA, K.D. Ferguson
8
and C.H. Brehautc
APiacer Pacific Limited, G.P.O. Box 4315, Sydney, NSW 2001.
8
Piacer Dome Canada Ltd, P.O. Box 49305 Bentall Station Vancouver, Canada V7X1 P1.
cPiacer Dome Inc, P.O. Box 350, Toronto-Dominion, Ontario, Canada M5K 1 N3.
ABSTRACT
Acid drainage (AMO or ARD) can have significant long term and financial implications for mining
companies and, in some cases, the community through abandoned minesites. Recent examples
highlighting the scale and nature of these implications are presented. This potential liability
and, in most instances, a 'fear ofthe unknown' has resulted in the demand by regulatory agencies
for overly conservative control strategies and significant performance bonds in an attempt to
provide assurances that any acid drainage problems can be managed, regardless of the risk.
Acid drainage predictions usually focus on assessing the probability that samples and waste
units will generate contaminated leachate. The rate of acid generation, migration potential and
possible consequences of acid drainage are usually considered in far less detail. Such analyses
are deficient and do not fully assess the risk of acid drainage. Examples of the application of
risk assessment techniques to acid drainage for all phases of mine development from exploration
through to mine closure are discussed. These risk assessment tools need to be applied more
consistently by operators and regulators to facilitate effective decision making at all stages of
a mining project.
1.0 INTRODUCTION
Environmental legacies associated with acid drainage, or Acid Rock Drainage (ARD) as it is
referred to by Placer Dome Inc. (PDI), at abandoned mine sites in North America have attracted
considerable media attention over the last few years. High profile 'disasters' such as Summitville
in Colorado have resulted in a public demand for more stringent government regulations and
standards with rehabilitation bonds (security deposits) based on actual site liabilities. As a result
of this adverse publicity and political pressure, regulators are now seeking overly conservative
ARD control strategies and performance bonds. However, the mining industry has been unable
to defend itself against these demands due to a lack of proven performance and successful
decommissioning of ARD sites. The trend is also for governments not to relinquish leases if
there is any doubt over the long term integrity of the rehabilitation strategy implemented, or to
request the establishment of reclamation trust funds to provide for the ongoing operation and
maintenance of ARD control strategies.
To date, assessments of the potential for ARD often focus primarily on the probability that the
waste will generate acid conditions. The migration potential is not usually considered in detail,
and the possible consequences of ARD release are examined only in broad terms. Such
assessments do not fully consider the risk of ARD. The consequences of ARD release depend
upon the nature of the ARD (strength and volume), the assimilative capacity of the receiving
environment, and the proximity and value of aquatic resources. In short not all impacts are
equal.
This paper discusses the significant long term and financial implications associated with ARD,
in two parts. Recent North American examples highlighting the scale and nature of these
implications are presented along with details of the Equity Silver mine where a reclamation trust
165 MURRAY, FERGUSON & BREHAUT
Liabilities Associated with AMO.
lund has been established. The second part of the paper addresses the application of risk
assessment techniques to ARD for all phases of mine development from exploration through
mine closure. These risk assessment tools provide the opportunity lor more effective decision
making by operators and regulators and more realistic estimates of actual ARD liabilities.
2.0 ARD LIABILITY ESTIMATES AND PROVISIONS
To indicate the scale and nature of the potential liabilities associated with ARD, some examples
of the estimated rehabilitation costs for North American abandoned sites and operating mines
are presented. Recent examples of the impact of these liabilities on the provision of performance
bonds are also discussed, with particular reference to the requirement for reclamation trust funds.
However, it should be noted that the major ARD liabilities in North America are associated with
the control of off-site impacts (e.g. water treatment prior to discharge).
2.1 Financial ARD Liabilities
In the USA, no estimate of the total liability attributable to existing and potential ARD problems
in the mining industry has been published. However, the MEND (Mine Environment Neutral
Drainage) program in Canada has estimated an ARD liability lor Canada's mining industry in
excess of CDN$5,000 million (Feasby and Jones 1994). This liability estimate has been the
main impetus behind Canada's co-operative approach (government and industry) in addressing
the ARD issue.
2.1.1 North American examples
The Summitville abandoned mine in Colorado has been declared a Superfund site by the US
EPA who estimated the total rehabilitation liability to be in excess of US$100 million. The US
EPA has taken over management of the site in order to minimise the off-site impacts associated
with ARD and are reportedly spending in the order of US$50,000 per day on containment and
treatment alone. The US mining industry has also contributed to the funding of studies to fully
define the problem and develop remedial strategies lor the site.
Similarly, Leadville in Colorado has been declared a Superfund site due to ARD impacts associated
with over a century of mining in the area. The State of Colorado estimated the clean-up costs
for Leadville to be in the order of US$290 million in 1986. Current lease holders in the area
are contributing to the control of ARD having spent US$27 million on a water treatment plant
and US$8.75 million on government response costs in 1993. In this case, some responsibility
for historic problems has been assumed by current operators in the area.
Sulfate contamination of groundwater at a large operation in Utah has attracted the attention
of US regulators who estimate the liability to be in the range US$500 million to US$1 ,200 million.
The operator of the mine is currently spending some US$35 million a year addressing the problem,
while the US EPA wants to declare it a Superfund site. The company believes it can ameliorate
the problem over time; however, should it be declared a Superfund site, it would be the first
operating mine in the USA to fall into this category. The result would be the company funding
the US EPA to manage the problem.
2.2 ARD Liability Provisions
As discussed previously, the potential liabilities associated with ARD have resulted in regulators
seeking financial assurances that rehabilitation strategies will be effective. Examples of the nature
and form of these securities follow.
MURRAY, FERGUSON & BREHAUT
166
liabilities Associated with AMD.
2.2.1 Performance bonds
At the Golden Sunlight mine in Montana USA, operated by POl, a progressive pertormance
bond of US$41 million is required by the State of Montana, the US Bureau of Land Management
and the US Bureau of Mines. As at the end of 1994, some $30 million had been provided to
the government as required by a pre-determined schedule. While Golden Sunlight is a relatively
dry site (evaporation exceeds precipitation), concerns regarding the scale of the waste dumps
and the potential for ARD have resulted in the demand for this substantial performance bond.
Field trials are being conducted in an attempt to demonstrate the adequacy of the proposed
strategy and thus reduce the security required.
The Equity Silver mine in British Columbia, Canada, also operated by PDI, was required to lodge
a reclamation bond of CDN$5.5 million, while modified waste dump rehabilitation strategy was
being implemented to ensure that these works were carried out. A post-reclamation trust lund
was also required and is discussed in the next section. Total reclamation expenditure on waste
dump rehabilitation including resloping and installation of a compacted clay layer following the
cessation of mining in early 1994 amounted to approximately CDN$12.5 million. The BC
government has released the security as the success of the waste dump rehabilitation strategy
is subject to the provision of the reclamation trust fund.
2.2.2 Reclamation trust fund
In 1981, the Equity Silver mine in BC detected ARD being produced by the waste rock dump
with associated downstream impacts. A program involving the collection of drainage from the
waste dumps was implemented and a water treatment plant constructed. The treatment plant
neutralises the ARD with lime which also precipitates the dissolved metals prior to discharging
the excess water to the environment.
Reviews of Equity Silver's proposed decommissioning plan between 1988 and 1990 highlighted
the long term requirement for treatment of ARD as the most significant issue yet to be resolved.
At the time of these reviews, lime demand and environmental operating expenditure were escalating
at a significant rate. Discussions regarding a long term reclamation bond were also proceeding
at this time and became more critical as the cessation of operations approached. It was soon
evident that no progress was being made in establishing long term trends and the potential cost
liability to the company.
In early 1991, a technical committee was formed to identify and resolve outstanding differences
between the groups. The mandate of the committee was to arrive at an unbiased probability
distribution of the expected long term post-closure operating costs. The committee was comprised
of representatives from Equity Silver, the Ministry of Environment, the Ministry of Mines and
an independent consultant who acted as a facilitator for the committee. Two specific rehabilitation
strategies were considered; the first employing the existing waste dump rehabilitation approach
of an uncompacted cover and the other involving dump resloping and placement of a 0.5 m
compacted clay layer under the cover.
This latter option had been tested over a 16 hectare site on the waste dump during 1990 which
suggested this strategy could significantly reduce treatment costs. Where the existing cover
design reduced infiltration to 40% of precipitation, the compacted clay layer reduced infiltration
to 5% although 1 0% was assumed for cost estimates. Models were used to predict ARD quantities
and qualities over time and thus lime consumption. Post-closure reclamation bond requirements
were calculated as the net present value ( 1 993 Canadian dollars) of annual operating costs
for a 100 year period using a 3% discount rate.
167 MURRAY, FERGUSON & BREHAUT
Liabilities Associated with AMD.
The post-closure reclamation bond estimates are presented in Table 1 lor a range of cases.
The reclamation bond size required varied from a low of CDN$22.8 million for the compacted
clay layer to a maximum of CDN$56.6 million should no additional rehabilitation work be undertaken.
On the basis of these estimates, it was concluded that a cover strategy incorporating a compacted
clay layer would significantly reduce the annual volumes of ARD generated and ultimately treatment
costs which would reduce the final reclamation bond required. On the merit of this work, the
BC Government requested a reclamation trust fund of CON $32 million be provided by Equity
Silver and commit to the compacted clay layer strategy for all waste dumps.
Equity Silver Mines Limited negotiated conditions associated with the reclamation trust fund
to enable the size of the fund to be reduced over time as the annual operating expenditure
decreases. The trust fund is only required to produce sufficient income to cover the annual
cost of operating and maintain the collection and treatment facilities. The lime demand to date
has peaked and is expected to reduce substantially now that installation of the clay layer and
waste dump rehabilitation have been completed. While this reclamation trust fund has been
implemented to address the ARD liability associated with the site, Equity Silver cannot relinquish
their liability (or responsibility) for the site.
Table 1. Reclamation Trust Fund estimates for equity silver.
Parameter Estimated Reclamation Bond Requirements (CDN$M)
Minimum Most Likely Maximum
Existing cover strategy
(40% Infiltration}
- Fixed costs
13.2 13.2 13.2
- Variable costs
4.1 4.1 4.1
-Lime costs
16.9 25.2 39.3
Total Cost
34.2 42.5 56.6
Clay Layer Strategy
(10% Infiltration}
- Fixed costs
13.2 13.2 13.2
- Variable costs
1.2 1.2 1.2
- Lime costs
8.4 11.5 17.7
Total Cost
22.8 25.9 32.1
The concept of reclamation trust funds is also being considered elsewhere for operations with
long term ARD liabilities. The Ontario Provincial Government is apparently proposing a post-closure
security for an operation in Ontario where significant ARD problems exist. In this case, a security
in the order of CDN$100 million has been suggested, although the justification for a bond of
this size is unknown. The tax liability of reclamation trust funds has recently been addressed
by a parliamentary committee in Canada involving submissions from the mining industry who
are proposing they be treated like retirement savings plans (superannuation funds) where tax
on income is deferred.
MURRAY, FERGUSON & BREHAUT 168
Liabilities Associated with AMD.
3.0 RISK ASSESSMENT TECHNIQUES AND APPLICATIONS
Evaluating and accepting risk is a necessary part of deciding to acquire or proceed with any
new mining project. Risk can be defined as the product of three questions:
(1} What can go wrong?
(2) How likely is it to go wrong?
(3} What are the consequences if it goes wrong?
Mathematically, risk is often defined as the probability of an event occurring multiplied by its
potential consequences. The result may be modified through the development of contingency
plans. In the context of ARD assessments, the above three questions could be interpreted as:
(1} Could ARD generation occur?
(2) What is the probability of ARD generation?
(3} What are the consequences of ARD generation?
The first question relates to possible sources of ARD, including ore and low grade stockpiles,
underground and pit walls, waste rock dumps, heap leach pads, tailing residues, road cuts, and
borrow pits. If a prevention strategy has been defined for the project, the possible failure modes
of the strategy may be examined in detail through a risk assessment. The probability of ARD
is usually examined in a geochemical testing program, e.g. acid/base accounting (ABA) and
kinetic testing. The ABA results are often compared with criteria, or the researcher may use
experience to estimate the probability of ARD generation.
ARO is not necessarily a concern if it does not migrate from the waste emplacement. This is
relevant in very arid climates or when materials with significant acid neutralisation capacity are
down gradient of the ARO generating material. Therefore, the probability of ARD generation
needs to consider both the capacity of the material to generate acidic or metal contaminated
drainage and the potential of any contaminated drainage to migrate from the waste emplacement
(consequences}.
Several methods of risk assessment are available that may be applicable to more completely
assess the risk of ARD. These include ranking schemes, pointscoring schemes, potential problem
analysis, failure mode and effect analysis (FMEA), fault and event trees, consequence evaluations,
modelling, and quantitative probabilistic analysis.
Simple qualitative risk assessments have been used by mining companies and regulatory agencies,
either intentionally or unintentionally, in assessing mining plans. The more sophisTicated techniques
have been used infrequently. FMEA was used for the assessment of two mining projects in
British Columbia (Pelletier and Dushnisky 1993; Van Zyl and Bamberg 1992}. Some techniques
being developed by POl and their applications are discussed in the following sections.
3.1 ARD Category Visualisation
As discussed by Murray et al. (1995}, in a paper presented at this workshop, POl have adopted
the NP/AP ratio as a screening criteria to access the risk of ARD from the data generated by
the predictive static tests (acid-base accounting). The distribution of samples within the different
NP/AP categories (ratio ranges} can be presented graphically to highlight the relative risk. However,
it should be noted that this approach does not take into account the potential consequences
of ARO generation and therefore should be applied within its limitations.
The NP/AP ratio may be considered a "safety factor" as used in other engineering analyses.
169
MURRAY, FERGUSON & BREHAUT
Liabilities Associalsd with AMD.
Higher safety factors are probably required for mines in wet environments where neutralising
(caibonate) minerals may be preferentially leached from the wastes relative to the rate of oxidation
of sulfide minerals. This is particularly relevant to waste rock materials where any samples
with a NP/AP ratio < 3 are usually subjected to further (kinetic) testing to assess the relative
reaction rates of the different constituents. A less conservative approach can usually be applied
to tailing materials as they are finer grained and more homogeneous than waste rock.
3.2 Issue Ranking Matrix
When evaluating new projects lor possible acquisition or development, relatively little information
is usually available to assess the risk of ARD. A simple approach is therefore required and
Fig. 1 presents an issue ranking matrix which has been developed by POl for this use. The
probability of ARD and parameter migration potential are ranked according to subjective low,
medium, and high ratings. The risk is categorised from 1 to 4 and can be modified by one level
depending upon the degree of environmental sensitivity (consequences).
Migration Potential
Low
Medium
High
Add one level for high environmental sensitivity
NOTES Deduct one level for low environmental sensitivity
Fig. 1. Issue ranking matrix for ARD potential.
The probability may be defined by; observation of ARD from existing facilities or outcrop seeps,
the presence of massive sulfides or carbonates in core, or the availability of mineralogical and
geochemical data. The parameter migration potential reflects climatic conditions, possible isolation
of sources (e.g. dry mine walls), or limited quantities of wastes (e.g. small waste rock dumps
in underground operations). Because the analysis is entirely subjective, the information and
assumptions used are generally explained in supporting documentation for future reference.
Also, the matrix can be re-evaluated and updated as more information becomes available, and
can be used for illustrating the relative differences between sites.
3.3 Fault Trees and Event Trees
The concept of examining possible 'faults' and resulting 'events' are integral parts of designing
a mine but have rarely been applied in a rigorous manner to ARD assessments. Fault trees
are typically used to identify all the mechanisms by which an undesirable event could occur.
The undesirable event is identified at the top of the tree, and all the subordinate events occur
in the lower tree structure. The technique is well suited to examining the probable success
MURRAY, FERGUSON & BREHAUT
170
Liabilities Associatad with AMD.
Failure of Flooded Tailing Impoundment
Fig. 2. Fault tree for failure of flooded tailing Impoundment.
of an ARD prevention plan. Figure 2 presents a fault tree for a plan to flood a tailing impoundment
to prevent ARD formation. Probabilities can be assigned to each event, the total probability
of failure calculated which identifies the most likely failure mode.
Event trees are used to examine the consequences of an initiating event in detail. For example
Fig. 3 examines the effect of an excessive storm causing an environmental release from a tailing
impoundment. The loading of contaminants could be estimated and possible impacts on the
environment determined based on the presence or absence of aquatic resources. These forms
of risk assessments can also be used to identify those potential events where contingency measures
are required.
Fish Not Present
'>J ? Fish Present
~
Fish Not Present
Excessive Storm
Fig. 3. Event tree for off-site release due to excessive storm.
171 MURRAY, FERGUSON & BREHAUT
Liabilities Associated with AMD.
3.4
Consequence Evaluation
Simple models can also be constructed to examine the consequences of ARD release. For
example, in some cases, the natural environment may have a significant capacity to assimilate
ARD. While this should not be relied on to control ARD, an evaluation of the possible consequences
may help to focus effort. An example of a POl application of consequence evaluation follows.
An open pit mine was proposed in an area of very high rainfall (> 4 m) with large streams containing
high alkalinity (both as dissolved bicarbonate and as calcite in sediments). In order to assess
whether ARD from the waste rock dump and open pit could cause a pH depression or problems
with heavy metal concentrations in the river downstream, a simple model was developed of
the following assumptions:
(1) The pH in the river may fall if the total alkalinity reaches zero;
(2} The total alkalinity is zero when the acidity load equals the alkalinity load;
(3} Acidity and alkalinity behave as conservative water quality parameters; and
(4) The open pit and waste rock dump would not generate ARD concurrently.
The critical acidity of the ARD from the waste rock dump and open pit was back-calculated using
a simple mass balance equation and average monthly flows.
( Fd X Ad} - (Fu X A)
As ~ ---=---
Fs
[1]
where; A.= acidity (negative alkalinity) of ARD (mg L-')
Fd =flow in river downstream of source (L sec-')
Ad = alkalinity in river downstream of source (assumed in this case to be 0 mg L-')
F" =flow in river upstream of source (L sec-')
Au= alkalinity in river upstream of source (mg L-')
F.= flow of acid water from sources (L sec')
Results indicated that the maximum acidity of the waste rock dump ARD that could be assimilated
by the receiving streams without a pH depression ranged from 1 ,600 mg L-' to 6,500 mg L-'
(as CaC0
3
}. Since ARD from other acid generating waste rock dumps of equivalent size around
the world has acidity concentration at these levels, it was concluded that, ARD generated by
the dump could be of concern. The dilution of ARD by the river ranged from ratios of 18:1 to
73:1 which are considered to be relatively low values, so the river could also be sensitive to
significant dissolved metal concentrations in any ARD.
The maximum acidity that could be tolerated from the open pit ranged from 5,500 mg L-' to
22,000 mg L-' (as CaC0
3
); which are relatively high values and unlikely to be realised in the
field unless the pitwall rock is extremely reactive. The dilution of pitwater by the river ranged
from ratios of 63:1 to 251 :1. Therefore, ARD from the pit is not as likely to cause a downstream
metal problem as drainage from the waste rock dump; hence the assessment and prevention
of ARD from the pit assumes a lower priority.
3.5 Probability Analysis
In quantitative analysis, a single discrete number in most instances does not adequately describe
risk as inputs to the assessment are uncertain. Probability analysis may assist in more fully
describing the actual risk. Probability analysis has been applied by Annandale and Chantler
(1992) to a mine site water and contaminant balance in order to estimate the probability of achieving
MURRAY, FERGUSON & BREHAUT 172
L1abil<ties Associated with AMD.
water quality parameters. The following case illustrates an application to estimate only the
probability of ARD.
For one particular mine, PDI assessed the probability that mass-weighted and surface-area-weighted
net neutralisation potential (NNP) for a waste rock dump would achieve selected criteria. The
analysis involved three steps; first, probability distribution functions were fitted to the NP and
AP datasets for each of five rock types with log-normal distributions providing the best fit. Second,
a spreadsheet model was constructed to combine the synthesised NNP distributions according
to the mass and the surface area of each rock type into single distributions for the entire waste
rock dump. One rock type was extremely friable and slaked to a much finer particle size than
the other four types and therefore had a greater contribution to the surface area weighted NNP.
Third, the distributions were sampled using a latin hypercube stratified sampling method to construct
the weighted NNP distributions.
The probability of the entire dump achieving selected criteria could then be determined directly
from the probability distributions. For example, the probability of the mixed waste rock dump
achieving a mass and surface area weighted NPP value less than zero (potentially acid generating)
were about 7% and 30%, respectively. However, criteria should not be the sole basis on which
predictions are made and the shape of the probability distribution functions also indicate the
range, central tendency, and moments of the data. In the case described here, the analysis
showed that the friable rock type caused a disproportionate increase in the acid generating potential
of the waste dump. Special material handling plans (subaqueous disposal) were developed
to address the higher potential for ARD from this material.
3.6 Modelling Consequences of ARD
Even where waste has an intrinsic capacity to generate ARD, acid drainage may not be generated
by the waste emplacement if sufficient neutralising minerals are present in the flow path. This
may be particularly relevant to tailing impoundments where carbonate minerals can occur in
the natural materials underlying the impoundment. Examples where modelling of ARD
consequences has been applied by PDI are outlined below.
The tailing residue contained an average of about 3.5% sulfur as pyrrhotite and pyrite and about
7% carbonate as calcite. Kinetic tests indicated the sulfide minerals were amenable to oxidation
and that in the long term, acid generation would occur if sufficient oxygen were present to support
oxidation. However, the phreatic surface (watertable) remained high in the deposited tailing
(maintaining a saturated state) and may limit the depth and degree of oxidation.
The model WATAIL (Scharer et al. 1993) was used to study the effect of various depths of the
phreatic surface on acid generation in the tailing. The model was applied to four separate areas
(nodes) of the impoundment. For node 1 the phreatic surface was assumed to be 4 m below
the surface; for node 2 it was assumed to be 2m and for nodes 3 and 4, 1 m below the surface.
Essentially the model examined the possible drawdown of the phreatic surface near one of the
impoundment walls.
The total depth of the tailing deposit was assumed to be 5 m for all nodes and the model simulated
100 years of oxidation and seepage. The generation of acid seepage did not occur for nodes
2 to 4 but was predicted to occur in 55 years for node 1 (Fig. 4). Even though the model indicated
acid seepage would not occur for some nodes, migration of those metals mobile at an alkaline
to neutral pH, such as zinc and cadmium, may still occur. Based on this analysis, it was decided
to increase the depth of tailing in the lower portion of the impoundment and to raise the phreatic
surface by constructing a water-retaining dam.
173 MURRAY, FERGUSON & BREHAUT
Liabilities Associated with AMD.
8.0
7.0
,
~
6.0
c:
::s -Nodel
:z: 5.0
.e:
--o-- Node2
:z:
4.0 c.
--+-- Node3
3.0
------Node 4
_)
2.0
0 20 40 60 80 100
Years In Simulation
Fig. 4. Predicted pH In seepage based on model simulation.
In a second example, a simple model was constructed to predict the sulfate and metal
concentrations in drainage from a waste rock dump where potentially acid-generating rock was
to be placed on top of a dump because of the mining sequence. A geochemical rather than
oxygen-limiting model was used since the quantity of potentially acid generating rock was small
and exhibited a low intrinsic oxidation rate in kinetic tests; calculations also indicated oxygen
was not limited, given the depth and reactivity of the material.
Results from a series of seven field kinetic (barrel) tests were used to determine oxidation rates
(sulfate and metal production rates). A mean sulfate production rate was normalised to the
percentage sulfur in the sample and the number of days between leachate sampling. The
normalised rate of sulfate production was found to occur over a fairly narrow range (highest
value two to three times the lowest value). An exponential decay rate was applied (time
0

5
)
to account for the build up of coatings during long dry periods and the resulting decrease in
sulfate production. To estimate metal concentrations, zinc, copper, and cadmium were correlated
with sulfate for the field kinetic data with correlations considered adequate for this level of modelling.
Figure 5 presents the data for zinc, where relationships were established for three ranges: high,
medium, and low reactivity based on sulfate production.
The model was run to predict sulfate and metal concentrations in drainage from the waste rock
dump based on recorded precipitation during the period July 1992 to October 1993 which was
the duration of the field kinetic tests. The drainage from each type of waste (high, medium,
and low reactivity and no acid generation) was weighted according to the possible tonnage and
surface area of the dump. The drainage was then diluted by the receiving water according to
simple ratios of catchment areas. The model predictions for zinc are presented in Fig. 6 which
indicates that high concentrations of zinc may occur. While the model was very simple, the
results produced are adequate for assessing the potential risk on which management decisions
can be based. As the results of this exercise indicated that zinc concentrations could be significant
in the receiving environment, a system for the identification and selective handling of problematic
ARD material was developed. An evaluation of contingency strategies was also conducted to
develop a plan should monitoring of receiving water indicate a need for mitigative measures.
MURRAY, FERGUSON & BREHAUT 174
Liabilities Associated with AMD.
I . H3 +CP120 * CP121 CP122 X CP123 + CP124 ... CP12i]
1 , 0 0 0 ~ - - - - -
...
.------
c
.Q
(i1
I..
c
; Cll
(.)
c
0.1L_:

8 :
(.)
c
N
X
LOW REACTIVITY
+ + + +
0.001 L.,_ ____ _L_ _____ __L ____ ___jL__ __ ....J
10 100 1,000 10,000
Sulphate Concentration
Fig. 5. Correlation between zinc and sulfate concentrations In lyslmeter leachates.
4.0 CONCLUSION
The formal application of risk assessment to environmental analysis in mining is relatively new.
The assessment need not be complex to provide a useful insight into the probability and
consequences of ARD and the need or level of mitigative measures. Risk assessment techniques
such as those discussed above show much promise for analysis of ARD problems, particularly
since the assessment of ARD remains an inexact science. Some possible applications of risk
assessment techniques to the various phases of mine development are shown in Table 2. These
tools are applicable to both operators and regulators of mines.
With the increasing application of these risk assessment techniques, more rational decisions
in mine design, approval, and operation should be possible. This approach will assist in reducing
the financial and long term liability associated with ARD, and the need for overly conservative
performance bonds and reclamation trust funds. The development of case studies where potential
ARD problems are prevented through good management practices is needed to promote confidence
amongst the regulators and public, alike.
175
MURRAY, FERGUSON & BREHAUT
Liabilities Associated with AMD.
40
'a; 35
.s 30
c
.S!
co
25
....
20
'E
Q)
0
15
c:
0
(.)
10
0
c:
5
N
0
Jul-92 Oct-92 Jan-93
May-93 Aug-93 Nov-93
Model Output Date
Fig. 6. Predicted zinc concentration by model simulation.
Table 2. Possible application of selected risk assessment techniques to ARD.
Risk assessment Project phases
technique
Operators Operators/Regulators
Acquisition Pre- Feasibility/ Operation Closure
feasibility Design
Ranking Matrix

Category

Visualisation
Fault/Event Trees -

Consequence

++
Evaluation
Probability - -

Analysis
Modelling - -

~
probably not applicable
+ possibly applicable
+ possibly applicable but may not have sufficient definition
very applicable
4.0 REFERENCES
Annandale, G.W., and Chantler, A.G. (1992). Uncertainty in planning ARD control measures:
Risk modelling and interpretation. In 'Risk AssessmenVManagement Issues in the
Environmental Planning of Mines'. (Society for Mining, Metallurgy, and Exploration,
MURRAY, FERGUSON & BREHAUT 176
Liabilities Associated with AMO.
Inc.: Littleton, Colorado.)
Feasby, G., and Jones, R.K. (1994). Report of results of a workshop on mine reclamation, Toronto,
Ontario, March 10-11, 1994, Ottawa. (CAN MET and Mining Sector: Natural Resources
Canada.)
Murray, G.S.C., Robertson, J.D., and Ferguson, K.D. (1995). Defining the AMD problem. 1.
A corporate perspective. In 'Proceedings of the Second Australian Workshop on Acid
Mine Drainage, Charters Towers, Queensland, 28-31 March 1995'. (Eds. N.J. Grunden
and L.C. Bell). pp. 3-15. (Australian Centre for Minesite Rehabilitation Research: Brisbane,
Australia.)
Pelletier, C.A., and Dushnisky, K. (1993). Qualitative environmental risk assessment applied
to the proposed Windy Craggy Project. Proceedings of the Seventeenth Annual British
Columbia Mine Reclamation Symposium, May 4 to 7, 1993, Port Hardy, B.C.
Scharer, J.M., Annabele, W.K., and Nicholson, R.V. (1993). WATAIL 1.0 User's Manual: A Tailing
Basin Model to Evaluate Transient Water Quality of Acid Mine Drainage. (Falconbridge
Ltd. and MEND Ontario, Institute of Groundwater Research: University of Waterloo,
Waterloo, Ontario.)
Van Zyl, D., and Bamberg, S. (1992). Qualitative environmental risk assessment for mine
development. In 'Risk Assessment/Management Issues in the Environmental Planning
of Mines'. (Society for Mining, Metallurgy, and Exploration, Inc.: Littleton, Colorado.)
177 MURRAY, FERGUSON & BREHAUT
178
Second Australian Acid Mine Dra1nage Workshop (Eds N.J. Grundon and L.C. Ball), pp. 179-190
ASSESSMENT OF THE RUM JUNGLE STRATEGY
FOR ACID MINE DRAINAGE CONTROL
J.W. BennettA and M.D. Lawton
9
AEnvironmental Science Program, Australian Nuclear Science and Technology
Organisation, PMB 1, Menai, NSW 2234.
"Water Resources Division, Power and Water Authority, G.P.O. Box 1096, Darwin, NT 0801.
ABSTRACT
By the time it was abandoned in 1971, the Rum Jungle copper and uranium mine site was a
major source of acid mine drainage, impacting regional river and groundwater systems.
Rehabilitation works were carried out from 1983 to 1986 and monitoring from 1986 to 1993
demonstrated that rehabilitation had been successful in terms of the original design criteria.
The Commonwealth Government has subsequently funded a five--year research program to
enable generic information to be obtained about the effectiveness of the various acid mine
drainage control measures adopted at the site. The work aims to quantify key processes,
including oxygen fluxes through the covers on the waste rock dumps, pollution generation rates
in the dumps, chemica/loads from the dumps, and pollution transport mechanisms through and
from the site, including through the water-filled open-cuts. The program also includes the
monitoring of water chemistry in the East Finniss and Finniss Rivers downstream of the site
and a study of the use of biological indicators to quantify the ecological impact of the chemical
pollution on downstream receiving waters. The further five-year period provides the
opportunity to monitor the long-term integrity of the engineered earthworks on the site and the
viability of the revegetation.
1.0 BACKGROUND
Ore treatment at the Rum Jungle uranium and copper mine began in 1954 and continued until
shutdown in 1971. Pollution of the Finniss River system by the mining and milling operations
had become apparent in the 1960's. As well as direct contamination by the uncontained
release of tailings and process effluent, there was also overflow and seepage of polluted water
from the open-cuts and surface and groundwater pollution by drainage from the waste rock
dumps and a heap leach pile.
The major environmental impact of the site was a consequence of the production of acid and
heavy metal salts by the process of acid mine drainage occurring in the sulfidic materials. The
three largest waste rock dumps were estimated to contain between 0.9% and 3% sulfur, mainly
as pyrite. Table 1 presents a summary of results of an extensive environmental survey carried
out in 1973-7 4 to quantify the annual release of heavy metals from each of the sources at the
mine site (Davy 1975). The location of these entities is shown in Fig. 1. The characteristics
of the waste rock dumps at this time are given in Table 2.
The 1973-74 survey also found that pollutant releases affected aquatic species in the river
system, with an almost complete absence of flora and fauna species for 8.5 km down the East
Finniss River, to the confluence with the Finniss River, and a reduced biodiversity in the
Finniss River evident for at least a further 15 km.
179
BENNETT & LAWTON
Rum Jungle Strategy for AMD ControL
Table 1. Annual release of heavy metals from each source at the Rum Jungle site
(1973-74).
Source Annual release (tonnes)
Copper heap leach pile
White's waste rock dump
White's open-cut
Intermediate waste rock dump
Intermediate open-cut
Dyson's waste rock dump
Dyson's open-cut
Tailings area
Old acid dam
Total
BORROW
AREA 1 /
\ .
\
-/;.3<,
ij/ '
'<' /)
~ / 60RAOWAREA6
~
. : ~ J ,..,..
Cu
32-42
29-53
8
16-30
3
0.2
1
5
-
95-142
'-y
\
I
Fig. 1. Rum Jungle mine site plan prior to rehabilitation.
BENNETT & LAWTON
180
Mn Zn
11-19 17-31
30
2.5-4.5 13-25
3 0.3
5
3
3.5
12
70-80 30-56
'
i
I'
'
~
N
I
0 500 1000m
SCALE
Table 2. Characteristics of waste rock dumps
before rehabilitation.
Parameter White's Intermediate Dyson's
Area (ha) 26 6.9 8.4
Volume (m
3
) 4x10
6
0.8x10
6
1.2x10
6
Mass (Mt) 8 1.6 2.3
Sulfur (g kg-') 32.7 30.6
Copper (g kg-') 0.86 2.0
2.0 REHABILITATION WORKS
Rum Jungle Strategy for AM 0 Control.
Furthermore, 100 km
2
of the
Finniss River floodplain were
affected to some extent by
heavy metal contamination.
The Davy report (Davey 1975)
made it clear that the waste rock
dumps were likely to remain
sources of pollution in the long
term, and proposed that the site
be rehabilitated.
In 1980 the Commonwealth Government made the decision to rehabilitate Rum Jungle, with
the Northern Territory Government to carry out the work. A detailed engineering report was
completed in 1982 and a Memorandum of Agreement was signed in 1983 by the
Commonwealth and the Northern Territory for the work to proceed. The objectives of the
rehabilitation works, as defined in the Agreement, were to achieve:
(1) A major reduction in pollution in water courses feeding the East Branch of the
Finniss River and in particular the reduction of the average annual releases of
copper, zinc and manganese in that river by 70%, 70% and 56% respectively,
as measured at the junction of that river with the Finniss River, in comparison to
the loads anticipated on the basis of the 1969-74 monitoring data;
(2) A reduction in public health hazards and in particular reduction of radiation levels
at the site at least to the standards set out in the Code of Practice on Radiation
Protection in the Mining and Milling of Radioactive Ores published by the
Australian Government Publishing Service in 1980;
(3) A reduction of pollution in the water contained in the open-cuts known as White's
and Intermediate; and
(4) Aesthetic improvements including revegetation.
The principal features of the rehabilitated program which was implemented between 1983 and
1986 are summarised below:
(1) White's, Intermediate and Dyson's waste rock dumps were reshaped, covers with
low hydraulic conductivity were constructed, drainage was constructed and the
dumps were revegetated. White's North dump was moved to the base of White's
dump before the latter was rehabilitated. The area originally occupied by White's
North waste rock dump was limed, covered with topsoil and revegetated;
(2) Tailings from the tailings dam and the low grade copper ore from the copper
heap leach pile were moved and placed in Dyson's open-cut. Ail three areas
were then covered, installed with drainage and revegetated;
(3) Water in White's and Intermediate open-cuts was treated using a water treatment
181
BENNETT & LAWTON
Rum Jungle Strategy for AMD Control.
plant based on a hydroxide precipitation process and then the East Branch of the
Finniss River was diverted through the open-cuts. The resultant precipitate was
buried on site; and
(4) Stream beds were cleaned up and the area was generally cleaned up.
The rehabilitation program has been fully documented by Allen and Verhoeven (1986). The
total cost of the work in 1986 dollars was $18.6 million; component costs are summarised in
Table 3. It is interesting to note that the cost of rehabilitating the waste rock dumps was some
$68,500 ha-
1
, including reshaping of the dumps before covering.
Table 3. Cost of components of
rehabilitation work.
Component Expenditure
1986 $(Aust)
Copper heap leach pile/tailings 3,685,925
dam/Dyson's open cut
Waste rock dumps 2,826,997
Open cut water treatment 6,234,221
Other areas 904,720
Management 4,666,040
Monitoring 322,346
Total 18,640,249
3.0 MONITORING TO 1993
Monitoring was recognised at the time
of the original agreement as being
important in demonstrating the
effectiveness of the rehabilitation
engineering works and was
incorporated as an integral part of the
program. There has been continuous
monitoring of a range of parameters
since 1986, and in a few cases the
monitoring was begun well before
rehabilitation started.
Statutory responsibility for the
management of the site passed to the
Conservation Commission of the
Northern Territory (CCNT) and the
monitoring program has been coordinated through the Rum Jungle Monitoring Committee,
chaired by a representative of the CCNT.
Through the Monitoring Committee, the Power and Water Authority (PAWA) of the Northern
Territory has been responsible for monitoring the surface water quality and hydrology in
waterways on the site, in the open-cuts and in the Finn iss River system, the Australian Nuclear
Science and Technology Organisation (ANSTO) (formerly the Australian Atomic Energy
Commission) has monitored chemical activity and water balance in the waste rock dumps,
groundwater hydrology and the ecology of receiving waters, and the CCNT has been
responsible for site maintenance and revegetation.
A report produced by the CCNT (Kraatz and Applegate 1992) described the results of
monitoring from 1986 to 1988. Another report covering monitoring from 1989 to 1993 is in the
final stages of preparation. What follows in this section is a summary of conclusions and
recommendations from the monitoring carried out to 1993.
3.1 Chemical Activity and Water Balance of the Overburden Heaps
Collection lysimeters were used to measure water infiltration through the covers on White's and
Intermediate waste rock dump. Infiltration through White's cover remained fairly steady since
the cover was completed in 1984, the average over the period 1988-1993 having been 2.2%
of incident rainfall, compared with 2.1% from 1984 to 1988. These figures are well below the
BENNETT & l.AWTON
182
Rum Jungle Strategy for AMD Control.
specified design value of 5%. The interpretation of data from Intermediate dump has been
made difficult due to the progressive failure of a number of lysimeters due to incorrect
installation, but it is most likely that the cover has continued to perform to specification.
Temperature measurements have indicated that both White's and Intermediate dumps
continued to cool since rehabilitation. In the five years to 1993 the maximum temperature at
the base of each dump had fallen by about 4 oc to around 34 C, with the rate of decrease
slowing with time as the temperatures approach lower values. This cooling is qualitatively
consistent with the oxidation rate having been greatly reduced by the rehabilitated works.
Measured pore gas oxygen profiles have shown generally lower concentrations since
rehabilitation. This is consistent with the covers acting to limit the rate of transport of oxygen
into the dumps and hence to limit the oxidation and pollution generation rates. In Intermediate
the concentration contours have not shown much variation with time of day or season,
indicating that the transport of oxygen into the dump has been controlled by diffusion and that
the diffusion coefficient of the cover remained fairly constant.
In White's dump the oxygen contours were more dynamic, with data indicating that there was
a greater penetration of oxygen into the dump in the dry season. It seems likely that the effect
is related to moisture content changes in the cover, such that as the cover dries after the end
of the wet season the gas-filled porosity increases, leading to higher gas diffusion coefficients
and a consequent increase in oxygen flux through the cover. This suggests that the cover
system on White's may not have been as effective in limiting the ingress of oxygen as it could
have been.
3.2 Groundwater Hydrology
Based on mathematical modelling of water and pollutant transport through waste rock dumps,
there would be a timescale of about 20 years for leaching soluble contaminants already
generated in White's dump before rehabilitation to the groundwater system. There has been
a measured decrease in rainfall infiltration through the dump due to the cover system, so over
a period of about 20 years the pollutant load leaving the dump could be expected to decline.
Such a decrease in pollutant load would represent a transient phase arising from the slow
response time of the system, but eventually a new equilibrium would be reached in which the
average load equals the pollution production rate. In the long term, any decline in the pollutant
load as a result of covering the dump must depend on a reduction in the pyritic oxidation rate
in the dump.
This picture is supported by field measurements which have shown that the concentration of
pollutants in the vicinity of White's dump did not change significantly in the five years to 1993.
The observations could alternatively be explained, however, by the presence of a large
reservoir of polluted water below the dump, which would respond only slowly to changes in
pollutant input rates.
3.3 Water Quality and Mixing Regimes of the Open,Cuts
The flooded White's open-cut was only about 50 m deep at mine closure, with a surface
diameter of approximately 360 m. Intermediate open-cut, although of similar 50 m depth, was
significantly smaller in volume with an approximate surface diameter of 210m to 270m.
Detailed physico-chemical measurements of the water column in both White's and Intermediate
open-cut water bodies were made prior to rehabilitation. These profiles have continued to be
183 BENNETT & LAWTON
Rum Jungle Strategy for AM D Control.
measured as an important element of the monitoring program since rehabilitation. Data from
the profiles have allowed the initial water treatment process to be closely monitored and the
effectiveness of the annual flushing of the open-cuts during the wet season flows to be
determined (Kraatz and Applegate 1992). With speculation concerning the source of pollutants
leaving the open-cuts and their contribution to the overall pollutant loads from the mine site,
a more detailed and systematic monitoring program was implemented in March 1992.
In White's, the initial water treatment in 1985 provided a lower density treated water layer of
20 m depth overlying the denser, untreated and highly polluted lower waters. A transitional
mixing layer was rapidly established and became a feature of the water column. Subsequent
flushing over six wet seasons resulted in the lower density surface layer deepening to about
27 m by mid 1993 and the transitional zone being almost totally eroded.
In 1993 White's had a reservoir of pollutants at depth, with copper concentrations of 60 mg L-'
in an acidic matrix, whereas Intermediate and minimal acidity throughout the profile and copper
concentrations decreased in the dense anoxic water below the pycnocline. Although the mixed
upper layer in Intermediate contained heavy metals, this pollution was a function, in part at
least, of the overflow water quality from White's which, in turn, depended on the degree of
flushing the water bodies received from wet season flow from the upper East Finniss River.
Temperature and conductivity profiles during inflow clearly demonstrated the influence of
vertical turbulence to depths of almost 30 m, well below the expected depth (Allen and
Verhoeven, 1986).
Measurements have shown that almost all the copper in White's above 32 m AHD can be
transported to the East Finniss River during a wet season, and that this can contribute around
half of the total annual copper load being carried from the site. During the following dry
season, prevailing dry south-easterly air flow, coupled with cooler conditions, can provide
conditions conducive to mixing of the White's water body to a depth of some 27 m. This
mixing is a mechanism whereby pollutants from the dense, heavily polluted waters can move
to the upper surface mixed layer. Diffusion from the lower waters and/or pyritic oxidation by
oxygenated surface waters on the mineralised walls of the pit have been found to be
insignificant mechanisms for pollutant transport.
3.4 Water Quality and Surface Hydrology
The principal aim of the work was to quantify loads of heavy metals at a point 5.6 km
downstream of the site on the East Branch of the Finniss River to check that the designed
reductions had been achieved. Further studies of water quality and surface hydrology on the
site and downstream of it were also undertaken to describe the pollutant transport mechanisms
and pathways.
Data on water quality and hydrology in the East Finniss was collected at gauging stations
GS8150097 (5.6 km downstream of the site) and GS8150200 (at the site boundary) (Fig. 2).
Data was also collected from a survey of waters downstream of the mine site along the East
Branch. Apart from the open-cuts, collection of data from within the mine site was largely
curtailed after 1988 and was halted completely after 1991. The results of the studies are
summarised below.
3.4.1 East Finniss River
Annual metal and sulfate loads leaving the mine site were calculated from data collected at
gauging station GS8150097. These results, together with historical data, are presented in
BENNETT & LAWTON
184
'1.' 1\_
'\..' "\ \{_/,." '
. " (
GS 1150097

\--I---
I _,fl.,... -.
'- I '

q

0 , lo;m
'--- J I
I
I
I
I
I

I-
I',
--1,
I''
\\
\'.
;' I /
.... ' 1.'
.v 1-
/ .,
I
'
Rum Jungle Strategy for AMD Control.
LEGEND
GS 11S0200 r- GAl,IG.ti() STATIOH
STAUL!
/
I


0
. '
,,_ .
, "
.I o ...
- ,_
Fig. 2. Location of gauging stations on the East Flnnlss River.
Table 4. They indicate that the rehabilitation has been successful in achieving the required
reduction In metal loads. As expected though, pollution continues to be discharged from the
site.
Concentrations of copper, manganese and zinc in the East Branch immediately below the mine
site were measured in 1993 to be two to three orders of magnitude above those measured at
control sites upstream of the site and were generally much higher than recommended for
drinking water and for the protection of freshwater ecosystem. Metal concentrations declined
with distance downstream of the mine site.
3.4.2 1991/92 Cumulative heavy metal loads at GS8150097 and GS8150200
In 1991/92 flow-weighted composite samples were collected at gauging stations GS8150200
and GS8150097 to check that loads at GS8150097 provided good measures of the pollution
loads leaving the mine site and that heavy metals were not precipitating along the 5.6 km of
river between the two stations.
Good agreement was found between the stations for the copper and zinc loads. In the case
of manganese and sulfate, however, there was an increase in the loads (from 6.5 I to 8.9 t Mn
and from 9,160 t to 12,600 t S0
4
). No surface flows were found to be entering the river
between the two stations but it has been suggested that groundwater flow discharging into the
East Branch through springs provides the additional source of manganese and sulfate.
Such groundwater discharge could be an important hydrological factor in determining the water
chemistry of recessional flows and residual water holes in this section of the river.
185
BENNETT & LAWTON
Rum Jungle Strategy ior AMD Control.
Table 4. Historical load data (tonnes) from GSB150097 together with a record of flow
and rainfall.
-----
Year Flow Rainfall Copper Zinc Manganese Sulfate
(x10
6
m
3
) (mm) total total total
--------
1969/70 '7 896 44 46 3300
1970/71 33 1611 77 24 110 12000
1971/72 31 1542 77 24 84 6600
1972/73 22 1545 67 22 n 5500
1973/74 69 2000 106 30 87 13000
1982/83 9.5 1121 23 5 6 1520
1983/84 48 1704 28 9 21 3600
1984/85 11.7 1136 9.1 4.1 '7.2 1600
1985/86 11.4 1185 3.7 2.7 8.2 4400
1986/87 13.2 1222 5.6 2.7 8.6 2870
1987/88 6.3 1064 3.2 2 5.4 1230
1988/89 35 1600 5.4 4.4 19.2 3940
1989/90 3.1 900 1.8 1.6 3.9 760
1990/91 40.5 1590 14.9 7.4 30.5 4000
1991/92 7.1 1002 3.8 2.7 9.1 1260
1992/93 29.9 1421 11.9 3.9 24.7 2696
3.5 Macro-Invertebrate Ecology of Receiving Waters
In 1993 a survey of macro-invertebrates in the East Branch of the Finniss River repeated and
extended an earlier survey carried out in 1973/74.
Macro-invertebrate fauna was stili less diverse and less abundant in the lower reach of the
East Branch compared with control sites. Nevertheless, the total number of macro-invertebrate
families found in the East Branch were found to have increased in comparison with the pre-
remedial survey, indicating some recovery.
3.6 Site Maintenance and Revegetation
Overall, the covers have remained intact and have required minimal attention although die-
back has occurred on the infilled Dyson's open-cut.
Pasture remains vigorous on revegetated areas despite the cessation of regular fertilisation
and some impact from wildfire. Weed invasion is a significant problem and consumes more
resources than originally envisaged.
BENNETT & LAWTON
186
Rum Jungle Strategy for AMD Control.
Despite fencing and declaration as a Restricted Use Area, site security remains a problem and
has implications for the future monitoring program.
4.0 RESEARCH PROGRAM 1993 TO 1998
Following a recommendation early in 1993 from the Rum Jungle Monitoring Committee, an
agreement was signed between the Commonwealth and Northern Territory Governments for
a further five years of monitoring and maintenance. The research was to be carried out by the
same Territory and Commonwealth agencies as before.
The primary objective of the further work was to ensure that the goals set in the rehabilitated
program continue to be met in the medium to long term. In addition though, it had been
recognised that the understanding of many aspects of acid mine drainage had improved since
the rehabilitation scheme had been designed, particularly with respect to rate control and
transport mechanisms, and that an expanded program of study was warranted. There was an
appreciation of the importance of the Rum Jungle site as a national facility, offering a unique
opportunity to provide generic information to the mining industry on the effectiveness of various
acid mine drainage control measures and monitoring techniques.
A particular aim of the current 1993/98 program is to quantify many of the key parameters and
processes involved in the generation and transport of pollutants at the site. The main activities
included in the program, with total Commonwealth funding of $1.13 million, are listed below:
( 1) To quantify the effectiveness of the covers on White's and Intermediate waste
rock dumps;
(2) To quantify the pollution loads flowing from the base of the dumps;
(3) To quantify the pollution loads from Dyson's waste rock dump and assess the
effectiveness of its limited cover system if it is found to be a significant source;
(4) To quantify the time dependent loads of pollutants flowing from White's and
Intermediate open-cuts to the East Finniss River;
(5) To measure the water and sediment quality in the Finniss River system
downstream of the site; and
(6) To quantify the ecological impact in the river system of pollutants from the site.
A description of the broad components of the program is given in the following sections.
4.1 Effectiveness of Cover Systems
Since the primary objective of the rehabilitation of the waste rock dumps was to reduce the
pollution loads generated in them, the effectiveness of the cover systems should be assessed
by quantifying oxidation rates in these dumps, since these rates determine the pollution
generation rates and represent the source of pollutants on site.
Two approaches will be taken to determine oxidation rates. The first will be to analyse
measured temperature profiles to determine the heat source distribution giving rise to the
profiles and relate that to the heat generated in the oxidation reaction. The second will be to
analyse measured pore gas oxygen concentration profiles and use measured oxygen transport
properftes of the covers and bulk material to estimate the oxygen consumption rates.
4.2 Pollution Loads from Waste Rock Dumps
The current monitoring program aims to quantify pollution loads entering the groundwater
187
BENNETT & LAWTON
Rum Jungle Strategy for AMD Control.
system from the waste rock dumps and to determine the relative contribution of each dump
to the total load in the East Finniss River. These loads will be correlated with estimates of
past and present pollution generation rates in the dumps (described above), making
allowances for transport timescales determined from measured water infiltration rates, to test
the understanding of these processes in waste rock dumps.
It is intended that the following projects will be undertaken initially to achieve these aims and
to address a number of issues which arose from the previous monitoring (Section 3.2):
4.3
(1) make measurements of any stratification in the groundwater below and around
the waste rock dumps;
(2) attempt to use geophysical methods to establish the size and position of any
plumes of polluted water from the dumps;
(3) estimate the velocity of water in these plumes, possibly taking stratification into
account; and
(4) install field equipment to allow loads from the waste rock dumps to be monitored
in the long term.
Pollution Loads from Open-Cuts
Conjecture concerning the pollution regime within the open-cut precinct was aired in several
reports detailing the results of monitoring from 1986 to 1991. The PAWA Water Resources
Division has embarked upon a reinvigorated program of sampling and analysis of data to
define the wet season flushing mechanisms of the open-cut water bodies.
The program includes monthly profiling using a Hydroiab Surveyor II multi-parameter probe
(measuring depth, pH, temperature, conductivity, dissolved oxygen and ORP) coupled with a
submersible pump and discharge hose to the surface which allows for collection of samples
for chemical analysis, including analysis of heavy metals.
In addition, gauging stations have been installed at the inflow to White's open-cut and at the
outflow from Intermediate. These stations will allow pollutant loads entering and leaving the
open-cuts to be determined and to be reconciled with estimates made from analysis of the
open-cut profile data. A re-examination of the surrounding bore monitoring data including
levels and water quality will be made to establish the degree of interaction between the
groundwater and open-cut surface waters.
4.4 Water and Sediment Quality in the Finniss River
Hydrographic and water quality data will continue to be collected at gauging station
GS8150097 on the East Finniss River for the duration of the program. This will ensure
continuity of data for annual load estimates of the specified pollutants to be generated and for
the temporal distribution of pollutant concentrations to be charted. Gauging at GS8150200,
just downstream of the rehabilitated mine site, will be discontinued after wet season 1994/95.
An assessment of load estimate variations between GS8150097 and GS8150200 over the
period from 1991/92 to 1994/95 will be made.
In conjunction with the ANSTO biological impacts studies (see Section 4.5 below). water
quality surveys of residual flows in the East Finn iss will be made in April/May of each year to
assess the impact of low dilution seepages from the mine site on water chemistry and to
examine the possibility of contaminated groundwater discharging off-site to the East Finniss
River. A gauging station has also been reinstated on the Finniss River (GS8150204),
BENNETT & LAWTON
188
Rum Jungle Strategy for AMD Control.
downstream of the East Finniss confluence, to collect hydrologic and water quality data for the
biological impact studies in that reach of the Finniss River system.
4.5 River Ecology and Biomonitoring
There has been an increasing awareness of the role which biological monitoring can play in
establishing the impact on the environment of activities such as mining and in assessing the
effectiveness of any rehabilitation measures implemented.
The two projects outlined below will be undertaken in this monitoring period to provide
quantitative measures of the degree of ecological recovery that has taken place in the Finniss
River system since the Rum Jungle rehabilitation.
4.5.1 Macro-Invertebrate survey of the Finniss River
A study in the Finniss River system will investigate the annual cycles of the diversity and
abundance of benthic macro-invertebrates in relation to water chemistry. The study will identify
the time in the annual cycle when detriment takes place, as well as those water chemistries
that are responsible. This information may be used to predict the degree of further recovery
that could be expected following any further reductions in annual pollution loads.
4.5.2 Archival monitoring of metal levels in the Finniss River
it has been found recently that laminations in the shells of freshwater mussels can provide a
record of historical water concentrations of a variety of metals, including copper, zinc and
manganese. More significantly, they record the bio-available fraction of these metals, being
most relevant to the assessment of the toxicity of these metals.
Shells of the mussel Velesunio angasi will be collected from the Finn iss River and analysed
to obtain a record of changing concentrations of heavy metals in the river over the past three
or four decades (being the lifespan of the mussel). This offers the possibility of charting the
rate of improvement in the river following rehabilitation of the mine site.
4.6 Integrity of Earthworks and Vegetation
Monitoring work will continue in the development of an appropriate long-term strategy for
vegetation on the site. The structural integrity of earthworks on the site will also be closely
monitored.
Further studies will be carried out to determine the cause of an area of die-back on the
landform over Dyson's open-cut and to suggest possible remediation approaches.
Site security for the protection of the rehabilitation works remains a problem and will be
addressed in this period.
5.0 CONCLUSIONS
Monitoring at the Rum Jungle site to 1993 had the primary aim of testing the effectiveness of
the rehabilitation works in terms of the original design criteria and these were found to have
been met.
The current research program has the principal aim of quantifying pollution generation rates,
189
BENNETT & LAWTON
Rum .Jungle Strategy for AMO Control.
pollutant loads and transport mechanisms in, through and from the various entities on the site.
It is expected that the measurement techniques and improved understanding of mechanisms
and processes developed in this work, in conjunction with the biological monitoring studies and
monitoring of the continued viability of vegetation and landform stability, will provide information
of value to the broader mining industry in dealing with problems of acid mine drainage.
6.0
ACKNOWLEDGMENTS
We wish to acknowledge our colleagues on the Rum Jungle Monitoring Committee and within
the participating agencies, whose work we have freely drawn upon.
7.0 REFERENCES
Allen, C.G., and Verhoeven, T.J. (1986). The Rum Jungle rehabilitation project, final report.
Northern Territory Department of Mines and Energy for the Commonwealth
Department of Resources and Energy.
Davy, D.R. (1975). Rum Jungle environmental studies. Australian Atomic Energy Commission,
AAEC/E365.
Kraatz, M., and Applegate, R.J. (Eds) (1992). The Rum Jungle rehabilitation project
monitoring report 1986-88. CCNT Technical report number 51.
BENNETT & LAWTON
190
Second Austrahan Ac1d M1ne Drainage {Eds. N.J. Grundon and L.C. Bell), pp_ 191-192
SUMMARY OF GROUP DISCUSSION SESSION
Objectives of Session
The objectives of this 2-hour session were to identify future technology transfer and research
needs in the area of acid mine drainage.
Although it was originally the intention to have small group discussions with reporting back to
a plenary session, the consensus of the group was to stay together for the whole of this session.
The major areas identified for future attention were:
1. Research Needs
Assessment of the National Extent of AMD and the Associated Liability
Document the extent of the problem in the context of mining type and climatic zone.
Acid-generating Processes In Waste Rock Dumps and Tailings
Clarify effects of such factors as oxygen, temperature, water, salts and bacteria on intrinsic
oxidation rate of sulfides.
Identify relationship between pollution loads exiting system and oxidation rate.
Determine effect of waste rock dump characteristics (sulfide content, particle size
distribution, bulk density, preferred flow paths, water balance, geochemistry) and time
on pollution loads.
Identify minimum set of parameters of waste rock dump or tailings which need to be
measured so that processes can be monitored.
Calibrate Laboratory Methods against Field Performance
Oxidation kinetics.
Neutralisation kinetics.
Effectiveness of Dry Covers/Barriers
Need to assess performance of covers over time in different climates.
Identify instrumentation required to assess cover performance (lysimeters, water
content/potential).
Determine effect of vegetation on cover efficiency (cover protection, water balance, root
penetration, oxygen consumption, differential effects of grasses, shrubs and trees).
Role of Models
Determine their role in prediction of AMD and long-term performance evaluation.
2.
Technology Transfer Needs
Future Workshop
The next AMD Workshop should be held in the Northern Territory in June 1997 soon
after the Fourth International Conference on Acid Rock Drainage in Canada (30 May
6 June 1997).
191

Você também pode gostar