Está en la página 1de 405

Mine Water Hydrogeology
and Geochemistry

Geological Society Special Publications
Society Book Editors
A. J. FLEET (CHIEF EDITOR)
P. DOYLE
F. J. GREGORY
J. S. GRIFFITHS
A. J. HARTLEY
R. E. HOLDSWORTH
A. C. MORTON
N. S. ROBINS
M. S. STOKER
J. P. TURNER

Special Publication reviewing procedures

The Society makes every effort to ensure that the scientific and production quality of its books matches
that of its journals. Since 1997, all book proposals have been refereed by specialist reviewers as well as
by the Society's Books Editorial Committee. If the referees identify weaknesses in the proposal, these
must be addressed before the proposal is accepted.
Once the book is accepted, the Society has a team of Book Editors (listed above) who ensure that the
volume editors follow strict guidelines on refereeing and quality control. We insist that individual papers
can only be accepted after satisfactory review by two independent referees. The questions on the review
forms are similar to those for Journal of the Geological Society. The referees' forms and comments
must be available to the Society's Book Editors on request.
Although many of the books result from meetings, the editors are expected to commission papers that
were not presented at the meeting to ensure that the book provides a balanced coverage of the subject.
Being accepted for presentation at the meeting does not guarantee inclusion in the book.
Geological Society Special Publications are included in the ISI Science Citation Index, but they do
not have an impact factor, the latter being applicable only to journals.
More information about submitting a proposal and producing a Special Publication can be found on
the Society's web site: www.geolsoc.org.uk.

GEOLOGICAL SOCIETY SPECIAL PUBLICATION No. 198

Mine Water Hydrogeology
and Geochemistry

EDITED BY

P.L. YOUNGER
University of Newcastle upon Tyne, UK

and

N.S. ROBINS
British Geological Survey, UK

2002
Published by
The Geological Society
London

THE GEOLOGICAL SOCIETY

The Geological Society of London (GSL) was founded in 1807. It is the oldest national geological society in the
world and the largest in Europe. It was incorporated under Royal Charter in 1825 and is Registered Charity 210161.
The Society is the UK national learned and professional society for geology with a worldwide Fellowship (FGS) of
9000. The Society has the power to confer Chartered status on suitably qualified Fellows, and about 2000 of the Fel-
lowship carry the title (CGeol). Chartered Geologists may also obtain the equivalent European title, European Geol-
ogist (EurGeol). One fifth of the Society's fellowship resides outside the UK. To find out more about the Society,
log on to www.geolsoc.org.uk.
The Geological Society Publishing House (Bath, UK) produces the Society's international journals and books,
and acts as European distributor for selected publications of the American Association of Petroleum Geologists
(AAPG), the American Geological Institute (AGI), the Indonesian Petroleum Association (IPA), the Geological
Society of America (GSA), the Society for Sedimentary Geology (SEPM) and the Geologists' Association (GA).
Joint marketing agreements ensure that GSL Fellows may purchase these societies' publications at a discount. The
Society's online bookshop (accessible from www.geolsoc.org.uk) offers secure book purchasing with your credit or
debit card.
To find out about joining the Society and benefiting from substantial discounts on publications of GSL and other
societies worldwide, consult www.geolsoc.org.uk, or contact the Fellowship Department at: The Geological Society,
Burlington House, Piccadilly, London W1J OBG: Tel. +44 (0)20 7434 9944; Fax +44 (0)20 7439 8975: Email:
enquiries@geolsoc.org.uk.
For information about the Society's meetings, consult Events on www.geolsoc.org.uk. To find out more about the
Society's Corporate Affiliates Scheme, write to enquiries@geolsoc.org.uk.

Published by The Geological Society from: Distributors
The Geological Society Publishing House
Unit 7, Brassmill Enterprise Centre USA
Brassmill Lane AAPG Bookstore
Bath BA1 3JN, UK PO Box 979
(Orders'. Tel. +44 (0)1225 445046 Tulsa
Fax+44 (0)1225 442836) OK 74101-0979
Online bookshop: http://bookshop.geolsoc.org.uk USA
Orders: Tel. + 1 918 584-2555
Fax +1 918 560-2652
E-mail bookstore@aapg.org
The publishers make no representation, express or
implied, with regard to the accuracy of the information
contained in this book and cannot accept any legal
India
responsibility for any errors or omissions that may be
Affiliated East-West Press PVT Ltd
made.
G-l/16 Ansari Road, Daryaganj,
New Delhi 110002
India
©The Geological Society of London 2002. All rights Orders: Tel.+91 11 327-9113
reserved. No reproduction, copy or transmission of this Fax+91 11 326-0538
publication may be made without written permission. No E-mail affiliat@nda.vsnl.net.in
paragraph of this publication may be reproduced, copied
or transmitted save with the provisions of the Copyright
Licensing Agency, 90 Tottenham Court Road, London Japan
W1P 9HE. Users registered with the Copyright Kanda Book Trading Co.
Clearance Center, 27 Congress Street, Salem, MA Cityhouse Tama 204
01970, USA: the item-fee code for this publication is Tsurumaki 1-3-10
0305-8719/00/$15.00. Tama-shi
Tokyo 206-0034
Japan
British Library Cataloguing in Publication Data Orders: Tel. +81 (0)423 57-7650
A catalogue record for this book is available from the Fax+81 (0)42357-7651
British Library.

ISBN 1-86239-113-0
Typeset by Alden Bookset, UK
Printed by The Cromwell Press, Wiltshire, UK.

Contents

YOUNGER, P.L. & ROBINS, N.S. Challenges in the characterization and prediction of the 1
hydrogeology and geochemistry of mined ground
BOOTH , C.J. The effects of longwall coal mining on overlying aquifers 17
WOLKERSDORFER, C. Mine water tracing 47
WHITWORTH, K.R. The monitoring and modelling of mine water recovery in UK coalfields 61
DUMPLETON, S. Effects of longwall mining in the Selby Coalfield on the piezometry and aquifer 75
properties of the overlying Sherwood Sandstone
ADAMS, R. & YOUNGER, P.L. A physically based model of rebound in South Crofty tin mine, 89
Cornwall
ROBINS, N.S., DUMPLETON, S. & WALKER, J. Coalfield closure and environmental consequence 99
- the case in south Nottinghamshire
MCKELVEY, P., BEALE, G., TAYLOR, A., MANSELL, S., MIRA, B., VALDIVIA, C. & 107
HITCHCOCK, W. Depressurization of the north wall at the Escondida Copper Mine, Chile
KUMA, J.S., YOUNGER, P.L. & BOWELL, RJ. Hydrogeological framework for assessing the 121
possible environmental impacts of large-scale gold mines
BANWART, S.A., EVANS, K.A. & CROXFORD, S., Predicting mineral weathering rates at field 137
scale for mine water risk assessment
BOWELL, RJ. The hydrogeochemical dynamics of mine pit takes 159
SAALTINK, M.W., DOMENECH, C., AYORA, C. & CARRERA, J. Modelling the oxidation of 187
sulphides in an unsaturated soil
GANDY, C.J. & EVANS, K.A. Laboratory and numerical modelling studies of iron release from a 205
spoil heap in County Durham
BANKS, D., HOLDEN, W., AGUILAR, E., MENDEZ, C., KOLLER, D., ANDIA, Z., RODRIGUEZ, J., 215
SÆTHER, O.M., TORRICO, A., VENEROS, R. & FLORES, J. Contaminant source characterization of
the San Jose Mine, Oruro, Bolivia
NUTTALL, C.A. & YOUNGER, P.L. Secondary minerals in the abandoned mines of Nenthead, 241
Cumbria as sinks for pollutant metals
YOUNGER, P.L., The importance of pyritic roof strata in aquatic pollutant release from 251
abandoned mines in a major, oolitic, berthierine–chamosite–siderite iron ore field, Cleveland,
UK
LEBLANC, M., CASIOT, C., ELBAZ-POULICHET, F. & PERSONNE', C. Arsenic removal by oxidizing 267
bacteria in a heavily arsenic-contaminated acid mine drainage system (Carnoules, France)
BROWN, M.M.E., JONES, A.L., LEIGHFIELD, K.G. & Cox, S.J. Fingerprinting mine water in the 275
eastern sector of the South Wales Coalfield
BANKS, D., PARNACHEV, V.P., FRENGSTAD, B., HOLDEN, W., VEDERNIKOV, A.A. & 287
KANNACHUK, O.V. Alkaline mine drainage from metal sulphide and coal mines: examples from
Svalbard and Siberia
HATTINGH, R.P., PULLES, W., KRANTZ, R., PRETORIUS, C. & SWART, S. Assessment, prediction 297
and management of long-term post-closure water quality: a case study–Hlobane Colliery, South
Africa
NUTTALL, C.A., ADAMS, R. & YOUNGER, P.L. Integrated hydraulic–hydrogeochemical 315
assessment of flooded deep mine voids by test pumping at the Deerplay (Lancashire) and
Frances (Fife) Collieries

LOREDO, J., ORDONEZ, A. & PENDAS, F. Hydrogeological and geochemical interactions of 327
adjoining mercury and coal mine spoil heaps in the Morgao catchment (Mieres, NW Spain)
NEUMANN, I. & SAMI, K. Structural influence on plume migration from a tailings dam in the 337
West Rand, Republic of South Africa
JOHNSON, K.L. & YOUNGER, P.L., Hydrogeological and geochemical consequences of the 347
abandonment of Frazer's Grove carbonate-hosted Pb/Zn fluorspar mine, north Pennines, UK
HOLTON, D., KELLY, M. & BAKER, A. Paradise lost? Assessment of liabilities at a Uranium 365
mine in the Slovak Republic: Novoveska Huta
REES, S.B., BOWELL, RJ. & WISEMAN, I. Influence of mine hydrogeology on mine water 379
discharge chemistry

It is recommended that reference to all or part of this book should be made in one of the following
ways:
YOUNGER, P.L. & ROBINS, N.S. (eds) 2002. Mine Water Hydrogeology and Geochemistry. Geological
Society, London, Special Publications, 198.

ADAMS, R. & YOUNGER, P.L. 2002. A physically based model of rebound in South Crofty tin mine,
Cornwall In: YOUNGER, P.L. & ROBINS, N.S. (eds). Mine Water Hydrogeology and Geochemistry. Geo-
logical Society, London, Special Publications, 198, 89-97.

Referees

The Editors are grateful to the following people for their assistance with the reviewing of papers sub-
mitted to this Special Publication

Dr R Adams University of Newcastle, UK
Prof A C Aplin University of Newcastle, UK
Dr C Ayora CSIC, Barcelona, Spain
Mr D Banks Holymoor Consulting, Chesterfield, UK
Dr S A Banwart University of Sheffield, UK
Dr L C Batty University of Newcastle, UK
Dr J P Bloomfield British Geological Survey, Wallingford, UK
Prof C Booth Northern Illinois University, DeKalb, Illinois, USA
Dr R J Bowell SRK Consulting, Cardiff, UK
Mr M M E Brown Camborne School of Mines, University of Exeter, Redruth, UK
Dr S P Burke University of Sheffield, UK
Dr S Dumpleton British Geological Survey, Keyworth, UK
Dr K Evans University of Sheffield, UK
Prof G Fleming University of Strathclyde, UK
Dr R Fuge University of Wales, Aberystwyth, UK
Dr P J Hayes Water Management Consultants Ltd, Shrewsbury, UK
Dr J A Heathcote Entec UK Limited, Shrewsbury, UK
Dr A W Herbert Environmental Simulations International, Shrewsbury, UK
Mr D B Hughes University of Newcastle, UK
Dr K L Johnson University of Newcastle, UK
Dr D G Kinniburgh British Geological Survey, Wallingford, UK
Dr B A Klinck British Geological Survey, Keyworth, UK
Prof L Lovgren University of Umea, Sweden
Prof D A C Manning University of Newcastle, UK
Prof J Mather Royal Holloway, University of London, Egham, UK
Mr P McKelvey Water Management Consultants Ltd, Shrewsbury, UK
Dr A E Milodowski British Geological Survey, Keyworth, UK
Mr B D R Misstear Trinity College, Dublin, Ireland
Ms K L Morton K L M Consulting Services, Johannesburg, South Africa
Dr C A Nuttall University of Newcastle, UK
Dr D W Peach British Geological Survey, Wallingford, UK
Prof M Sauter Universitat Jena, Germany
Dr J M Sherwood Environment Agency, Exeter, UK
Dr P L Smedley British Geological Survey, Wallingford, UK
Mr R G Smith Environment Agency, Nottingham, UK
Mrs J A Walker Scott Wilson Kirkpatrick, Chesterfield, UK
Mr D C Watkins Camborne School of Mines, University of Exeter, Redruth, UK
Dr N R Walton University of Portsmouth, UK
Dr R S Ward Environment Agency, Solihull, UK
Dr C Wolkersdorfer Technische Universitat Bergakademie Freiberg, Germany

.

198. such as disposal of high. in which relatively modest investments in hydrogeochemical control measures during the exploration and exploitation phases of the mine life cycle will yield dividends in the post-closure phase. the conjunctive application of physical and geochemical evaluations will eventually provide much-needed predictive tools to inform site management decisions in the future. The post-closure studies described in this paper and in this volume (as well as elsewhere) highlight the dearth of hydrological data that are usually available when compared with the wealth of geometric information available from mine abandonment plans. Newcastle upon Tyne NE1 7RU. at the global scale the ization of active and abandoned mine systems. Special Publications. One of the principal legacies of past mining in Britain is water pollution emanating from abandoned mine voids and waste rock depositories.L. Underground a future renaissance in deep mining is to be flint mines of Neolithic age are still accessible in expected within the next few decades. Maclean Building. and tories. ROBINS2 1 Hydrogeochemical Engineering Research and Outreach. Perhaps the single most important lesson is that there can never be too much information on mine hydrogeology and geochemistry available at the post-closure phase. This has necessitated many expensive technical evaluations and remedial programmes in recent years. at present and to identify challenges for future research. mainly in the form of open-pit mining.00 © The Geological Society of London 2002. it should be given appropriate consideration from the outset. p. & ROBINS. University of Newcastle. UK Abstract: Although mining is no longer a key industry in the UK. 0305-8719/02/$15. 1) and North America (Bowell 2002) provides valuable lessons for the mining learned the hard way industry worldwide.the hydrogeological and geochemical character- worked areas of Europe. and will indeed prove southern Europe (Shepherd 1993). (eds) 2002. It is the purpose of this paper to outline mining has continued apace ever since. London. L.S. the international mining industry continues to expand. and by the lessons to be learned from recently closed deep Bronze Age underground metal mines were well mines in Europe will retain applicability else- established in many parts of western and where for decades to come. With such measures in place. and enhanced monitoring data to hand. Department of Civil Engineering. 1 Geological Society. Mining invaluable in the development of other uses of activity accelerated with the dawn of the underground voids. N.Challenges in the characterization and prediction of the hydrogeology and geochemistry of mined ground P. It is advocated that the collection of appropriate environmental data is built into the initial mine development plan and that monitoring commences from the green field site onwards. . Exploitation of the Earth's mineral wealth by (which often reflect ground water-related slope mining has been an unrelenting pursuit of stability constraints. from which important lessons may be drawn for the still-growing mining industry overseas. the southeast England (Holgate 1991). The recent closure of some very large surface mines The British experience: some lessons in Europe (Fig. With many of the larger Even the newest mining economies will even- open pits currently nearing their feasible limits tually come to share the superannuated status From: YOUNGER. whilst those of predicting mine water quality are even greater. YOUNGER1 & N. Wallingford Oxfordshire OX 10 8BB. Hence. The uncertainties related to predictive modelling of mine water arisings are considerable. and a call for 'defensive mine planning' is made. mining industry continues to expand. UK 2 British Geological Survey. 2002). Although some of the key findings of recent research into mining has long been in decline in the much. mankind since the earliest days.and Industrial Revolution when coal came to occupy medium-level radioactive wastes in deep reposi- a central place in the resource inventory. 1-16. see McKelvey et al. Mine Water Hydrogeology and Geochemistry. Numerous pointers towards robust mine water management strategies are identified. S. As this phase is also the longest in the overall life cycle of any mine.

When large systems of inter. 1. showing the pit lakes that have developed in the mine voids since coaling ceased. through which considerable quantities of groundwater entered the mine. used with permission. YOUNGER & N.) which now characterizes much of the European the former mining districts of Europe should mining sector. ROBINS Fig. 245 m in height.2 P. (Photograph: P Younger. strategies.) (b) The Westfield site viewed from approximately the same position on 12 December 2001. L. as explained in the text. (a) The main working pit (western basin) of the Westfield Opencast Coal Site. Scotland. Nowhere are the long-term environ- the experiences of the environmental conse. catalogue number D 1748. (Photograph: British Geological Survey. mental consequences of mining and mine closure quences of various closure strategies gained in more evident than in Britain. as it appeared in the height of its productivity in 1974 (after Robins 1990). prove invaluable to developing intelligent closure connected mines finally come to be abandoned. S. the outermost benches of which lie along the Ochil Fault. The far side of the pit is the highwall. Fife. from which some of .

Banks et al 1997. Staffordshire Yorkshire. la). the damage that past mining has done and is now most notably in North Wales. follow. Robins 1990. and Newcastle upon Tyne are testimony to the ensuring that Britain could therefore obtain industrial wealth created by the Victorian coal mined commodities from a wide range of mining industry. 1998. Ferruginous discharges from coal mines in UK (after Younger 2000a) Coalfields Number of Mean Fe Total length of rivers and identified concentrations streams contaminated discharges in discharges by ochre (mg/1) (km) South Wales. Copper activity. 2002.g. However. Brown et al 2002). and is therefore mental policy in favour of other sources of indulged in to some degree in the examples that energy (Powell 1993). by polluted discharges from abandoned mines • while it is. Nottinghamshire. North Wales 90 17 109 Warwickshire. Glasgow of the world until the early twentieth century. The costs associated with responsible minerals were expensive to transport and remediation of such pollution are often so were therefore largely sourced from home. there remains a substantial legacy of bings of West Lothian. Goldring & Juckes 2001). The magnificent Georgian architec- Although the British Empire dominated much ture of the city centres of Manchester. Younger resources the world over. CHARACTERIZATION AND PREDICTION OF MINED GROUND 3 the best-documented cases of mine water small price to pay (Fig. Large Roman and Mediaeval workings for volumes of Carboniferous grits and sandstones argentiferous galena are scattered through are contaminated with sulphate-rich ground- much of the uplands of Wales. and both doing to the aqueous environment. true that the British (about 400km from coal mines (Table 1). Less apparent to the casual observer is mining was underway in the Bronze Age. While hindsight is of little and the demise of the domestic coal value for the owners of existing problematic mining industry is ongoing even as we abandoned mine sites. for two reasons: Ml motorway between Nottingham and Shef- field (due to coal mine subsidence) and the • mining was carried out in the 'Cassiter. the uneven surface of the mining within the UK itself. compensated for any misgivings over polluted air 2001. with Empire was able to exploit mineral the remainder from metal mines. battered by a combination of cheap owners of other mines which are currently in the imports of coal and changes in govern. the heavier bulk 2000a). indeed. Hence. The benefits of full pollution are recorded (e. substantial that they beg the question as to It is only since the mid twentieth century whether they might have been minimized had our that the bulk of UK demand for ferrous forefathers pursued extraction and mine waste and base metal ores has been satisfied by management a little differently (e. One of the most problematic aspects of When mining was the major source of effective mine site management is the mismatch employment in much of northern and western between a keen twenty-first century environmen- Britain. and water.g. Younger 1994. it has some value for write.g. the unsightly oil shale sources. NRA employment and municipal prosperity more than 1994. Younger & imports (e. Harbourne 1995). planning or extractive phases. Current estimates are that some substantial environmental legacy from 600 km of British streams and rivers are damaged very ancient mines. 5 2. northern waters. there is a watercourses. Cumbria 25 20 35 Scotland–Midland Valley 167 27 180 . which cause serious ecological degra- England and the Isle of Man (with a few dation where they discharge to surface mines also in Scotland).5 Derbyshire. numerous flooded areas of former farmland in ides' (widely considered to refer to the Cheshire (due to solution mining of halite) are tin mining district of Cornwall) from the other less desirable consequences of past mining earliest eras of recorded history. environmental damage was considered a tal consciousness and the (often undocumented) Table 1. Durham and Northumberland 120 17 61 Lancashire.

The Wheal Jane case illustrates some inter- generation 'hotspots' within the bodies of mine esting points of wider interest. which was occasioned and surface waters from mining for a by incautious holing through the Ochil Fault in very long time. This pollution prompted quality. The years. making Westfield the deepest opencast Western Approaches of the English Channel coal mine ever worked in Europe (Grimshaw (Bowen et al 1998). only to close again in rock bund within the void. mineralized water crossing the Ochil Fault has the excessive costs of dewatering which been retained to form a pit lake of good quality. prevents the a priori identification of acid.) The sudden release of an esti- Figure la shows the main working pit as it mated 50 million litres (Ml) of highly polluted appeared in 1974. waste. even where a mineral deposit current after-use plans for the site are visionary has only been mined within the last few decades. S.2. inducing inflow of good quality methods and the gradual change in groundwater from Devonian strata which lie to environmental consciousness have worked the north of the site (Robins 1990). in the period of peak output. the relatively weakly 1978 for hydrogeological reasons (i. million at the time of writing.3 1250mgl– 1 (Robins 1990). Such problems are further exacerbated environmental quality and does not limit the where frequent changes in site management have scope for future generations to make their own resulted in the loss of all non-statutory records of decisions about the best use of the site. which remains mining has been underway for thousands of markedly acidic at the time of writing. obvious that this installed to control water levels and discharged problem is at its most marked in areas where water quality in the other pit lake. the unacceptability. Another problem arising from previous mining decisions is the unnecessarily high water • Although there has been a risk to ground make of the open pit. and peaked in the Victorian era. One of the few drawbacks of this water ranging between 8 and 15 Ml day– 1 has phenomenally rich site was its poor mine water continued ever since. with total iron concentrations reaching immediate public expenditure of some £8. Ib). unusually rich in coal: the site had an overburden The most infamous case of acid mine water to coal ratio of only 2. water and sludge from an old adit resulted in The highwall of the pit reached a total height of highly visible staining of the Fal estuary and the 245m. an underground in the 1950s. However. However. L.4 P. The Westfield OCCS worked a whose work pre-dated the emergence of the sequence of Carboniferous strata which was concept of sustainable development. and a further sum the principal source of such high iron concen. that is that the site is owners many years ago can act as serious res. which compares with pollution to have arisen in Britain within the last values in the range 15–30 at most other Scottish few decades is without doubt that associated with opencast sites. being developed in a manner which meets the traints on post-closure management and re-use of needs of the present which both improves a site. ('Wheal' is a Celtic (Cornish) word 25 million tonnes of coal (Grimshaw 1992). planners are unquestionably limited by the Scotland. provides a revealing example of these actions of previous generations of miners sorts of problems. The modern workings river system. of similar magnitude for the identification of a trations is active pyrite oxidation in and above long-term strategy to deal with the discharge the zone of water table fluctuation in the large (Knight Piesold & Partners Ltd 1995). YOUNGER & N. perhaps. state-of-the-art facilities have been or even centuries. a few months the absence of detailed records of mine waste before the 10th anniversary of the original disposal practices in the 1960s and early 1970s outburst. By judicious positioning of a major began in 1969. It is. A flow of contaminated 1992).e. the scale of modern mining the 1970s. between 1961 and 1986 it yielded more than Cornwall. and wholly in line with the modern philosophy of the legacy of mining decisions made by previous sustainable development. mining and mine waste disposal practices. The current site owners are 2000 BC. Westfield commenced production the 1991 closure of Wheal Jane. fell upon Wheal Jane after the adjoining . ROBINS realities of site evolution over periods of decades Meanwhile. Current masses of waste rock and washery fines that were total expenditure at this one site had reached £20 tipped into the pit some decades ago. and in its main productive period tin–zinc mine located in the Carnon Valley. Metal mining began in the former open pits have partly flooded to form pit area around Wheal Jane as long ago as lakes (Fig. Now that together to exacerbate risks to the point of coaling has finally ceased at Westfield. implementing a number of measures to prevent before ceasing altogether in the Carnon the site posing any problems to the adjoining Valley in 1913. It seems likely that million for amelioration works. meaning 'mine'. However. The the options open to today's engineers and Westfield Opencast Coal Site (OCCS) in Fife.

indus- gives rise to most concern and most trial or residential areas. (1999) as follows. constructed over former mine entries (e. on freshwater fish. Temporarily accelerated mine gas emis- consultation by the European Commission for a sions.g. driven ahead of the rising water new EU Mine Waste Directive. longer acceptable in a rapidly greening leading to acidic and/or ferruginous Carnon Valley? discharges (Table 1) (which sometimes • The environmental costs of mine water also contain elevated levels of other pollution are illustrated in bald cash terms ecotoxic metals. although there are no workings. methane is is wrought by polluted drainage from abandoned such a light gas that it will have normally workings than from waste rock/tailings deposits. primarily through slaking of . Overloading of sewers by mine water given rise to this debate has been fostered by two inflows. In this context. environmental damage that was once acceptable when mining was the major a. For 12 simple: even the largest spoil heap receives years the mine lay abandoned. 2001). can extend to abandonment of public water treat interventions. a survey of all known mine water pathway exists.g. The consequences of such pollution costs of remedial-preventative pump-and. Surface water pollution. or because the severity. Laudable as this table (e. Davis & Battersby 1985). the 1999). in Spain (Grimalt et al 1999) and at Baia Mare in e. shallow old failure runs the risk of overlooking the main workings. particularly remedial costs at the majority of other where structures have been inadvertantly mine water pollution sites in the UK. Robins & events. ishment of aquatic flora and fauna (Kelly mingly unexpected outburst at surface. there have been widespread calls for Younger 1996.g. g. and cases. more environmental damage flammable methane emissions. tighter regulation of mine waste. Localized flooding of agricultural. launching new regulatory legis. vented to surface anyway if an upward For instance. CHARACTERIZATION AND PREDICTION OF MINED GROUND 5 Mount Wellington Mine closed in April problem in the long run. The heightened consciousness that has d. The reason for this is 1978. Temporary loss of dilution for other whether a more specific regulatory framework is pollutants in surface waters as former needed to control pollution and related hazards pumped discharges cease to augment flows associated with catastrophic spills from mine (e. with spoil heaps and tailings repositories will thus lurk above the water table until contributing only about 2% of the total (Younger hydraulic ally forced upwards. and removal of a critical food table fluctuates in shallow workings has source for birds and mammals which prey been the main focus of remedial action. NRA represents an upper-bound estimate of the 1994). which is more dense than air. Younger These two observations have particular sig. which probably previously 'clean' streams (e. the more discharges in Scotland found that abandoned worrying prospect is of 'stythe' (carbon workings were responsible for around 98% of dioxide). most notably zinc) into by the Wheal Jane case. While media intention may be. attention is usually paid to the risk of in the longer term. nificance in the current Europe-wide debate on c. Is this because the spectacular 1991 order of decreasing incidence/environmental outbreak was truly unique. arranged in approximate tion. Younger & Adams 1999). it is nevertheless the case that. and this is precisely the process which b. 2002).g. Hence. The risk of subsidence as rising waters lation in response to a spectacular tailings dam weaken previously dry. resulting in f. poisoning of land animals that drink ongoing generation of acidity as the water the water. Following these two migration of mine water (e.g. open. Robinson 2000). at Aznalcollar precipitates (see Younger 2002). It is The full range of problems associated with the only the post-closure pollution following flooding of abandoned mine workings in Britain the 1980-1991 period of working that has has been catalogued by Younger & Adams finally given rise to environmental agita. records of any environmental outcry. fish deaths and impover- heralded by an instantaneous and see. exacerbated by clogging by ochre spectacular tailings dam failures. Younger & LaPierre 2000. and will recharge from a very small area compared to the certainly have been discharging acid water hundreds of square kilometres underlain by old to the Carnon River. Although its onset was supply intakes. after voids have source of employment in the area is no eventually flooded up to ground level. Pollution of overlying aquifers by upward Romania (UNEP 2000). Banks et al 1996). sites.

The et al. Nuttall et al on sound hydrogeological principles (e. On this basis. Holton et al. such as RBCA (risk. 1995). Younger (2002).6 P. etc. etc. Nevertheless. However. which is largely hydrogeological and and pose a health risk to rock climbers and geochemical in technical content (e. 2002). it is important to add amenity.g. preferably executed within a risk bulk of research to date has concerned the rate of assessment framework. Towards robust management strategies Then in 1996 the Environment Agency and Northumbrian Water Ltd launched a joint pro- On first inspection..g. Banwart & tions and flow rates. • damage to property and commerce. with geochemical work Malmstrom 2001. loss of sport fisheries. the specific problem of acidic pit lakes to the • the expenditure associated with various above list of issues. and it has only been since the advent of plumes of polluted water in the surrounding the UK Coal Authority in 1994 that a rolling aquifers down the hydraulic gradient (Neumann programme to address the worst of the abandoned & Sami 2002. the nece- flow (by evaporative water loss) or else they may ssary investment has not always been forth- be 'flow-through lakes'..g. where they may water pollution from mines is small compared be a principal mine water management issue with that of the potential damage and cost (Bowell 2002. examples of polluted pit lakes formed in former open-pit mine voids in the UK (the Westfield site It is frequently found that the cost of a mentioned above being a case in point).2002. Polluted pit lakes pose established. as advocated by Johnson & the desirability of alternative remedial actions.2002). Holton et a/. YOUNGER & N. they are preventative strategy in relation to post-closure widespread in other countries. Kuma 2002. damage Rebound in Abandoned Mineworkings.2002. this list of possible impacts gramme of research into 'improved modelling of of water level rise after mine closure (a process abandoned coalfields' (see Younger & Adams termed 'mine water rebound') is bewilderingly 1999 for details). L. finance to support independent scientific research into mine water hydrogeology and geochemistry was virtually non-existent before the mid-1990s. Although there are few remedial strategies. and the disin- provide roof support (e. The costs which need to be evaluated include Predictive physical modelling tools developed (Younger & Harbourne 1995): by UK-based researchers since the mid-1990s • environmental damage. funds for remedial actions have been scarce. Holton being as yet far less well developed. • damage to society. objective is the conjunction of investigations into tal cost-benefit analysis should be developed. No comparable programme yet exists environmental hazards of their own. A future et al 2002). and Authority have funded significant research into thus the key to effective management is to employ various aspects of the topic (see Banwart a characterization and prediction strategy based et fl/. Smith & Colls centive to urban redevelopment posed by 1996). a socia–-environmen. These pit of amelioration (Younger & Harbourne 1995. Knight others who may informally use the pit lake and Piesold & Partners Ltd.. Gandy & Evans 2002.g. Whitworth 2002). Banwart et al 2002. 2002). ROBINS seat-earths underlying pillars of coal left to to commercial fisheries. such as loss of have included: aquatic habitat and damaging the liveli- hoods of terrestrial animals that feed on • semi-distributed. which may mine water discharge is a source characterization accumulate in the closed depression of the pit initiative. visible stream pollution. such as for metalliferous mines. where surroundings for recreational purposes (e. both the physical and geochemical evaluation at which provides the principal tool for deciding on the same time. lumped parameter aquatic plants and animals. Subsequently. see Younger et al 2002 for further coal mine discharges in the UK has been discussion of this matter). Robins et 0/. mine water rise and prediction of outflow loca- based corrective action) (e. such as loss of visual Looking world-wide. the full range of impacts is Environment Research Council and the Coal unlikely to be experienced in any one case. such The best example of this genre is the as rendering further abstraction of water for Fortran code 'GRAM' (Groundwater public or industrial use impossible.g. exposing migratory waterfowl to toxins if they The foundation of a successful cost-benefit rest on the lake (Bowell 2002) and/or releasing evaluation of remedial options for a polluting harmful gases (such as radon). . models applicable at fairly large scales. the Natural broad. In reality. S. which give rise to coming. lakes can form sinks for regional groundwater Robins & Younger 1996).

Hydrogeological uncertainties ABA is generally far more difficult to apply to and challenges other settings with more complicated hydro- geological characteristics. in which a codes to a novel object-oriented code simulating pipe network model (representing major pyrite oxidation and the attenuation of acidity by mine roadways. few recorded instances: a good early example is which has been successfully linked to provided by Hattingh et al (2002). have been taken to couple numerical simulation • full physically based models. laboratory studies (e. Sherwood & 'stand-alone' mode by Banwart & Malmstrom Younger 1997. NET code. pollutant generation potential of strata under much remains to be done.g.) is routed through a a variety of reactions identified from field and variably saturated porous media (the VSS. Simpler water flow models coupled with digital . such as many base Uncertainties metal mines and virtually all deep mines (for coal Much of the discussion that follows derives from or metals). have found a certain amount of tentative Younger & Adams 1999). Simple assay techniques management strategies is to be realized. application in such cases (e.g. While these con- ditions may be satisfied to some degree in open- cast coal mining situations (Kleinmann 2000). Adams & tion protocol was developed for application Younger 2001). Although all the papers in with the content of buffering minerals (princi. Younger 2000b). various hydrological simulation codes It is clear that significant areas of uncertainty (Dumpleton et al 2001. While advances in this regard have been groundwater system tend to be topographic lows made for the case of pit lake wall rocks (Bowell (for surface decant) or stratigraphic/structural 2002) the geometric disposition of reactive lows (for decant to adjoining aquifers). provide some information on the these uncertainties and rise to these challenges. for real problem sites remains a state-of-the-art ware (VULCAN) that allows identification venture. and these must be addressed if the Predictive geochemical models have a much challenge of developing robust mine water weaker pedigree to date. A post-processing water quality predic- Burke & Younger 2000. such as those belonging to the PHREEQE cal mapping (e. but severe problems of parameterization using distributed. The code and its mass-balance models have been coupled to application has been widely discussed GRAM by Sherwood (1997). Banwart & Malmstrom Obviously. declines in head-dependent inflow as Although this has proved to be fairly robust in rebound proceeds. Younger & Adams 1999. and the need for an alternative to simple adjoining aquifers? ABA is imperative (e. There are minerals in disturbed strata surrounding deep two basic approaches to identifying such features: mine voids is far harder to characterize/ conceptualize. Robins et al 2002) remain with regard to hydrogeological and including GRAM. such following section we attempt to identify some of as 'acid-base accounting' (ABA) in which the the key hydrogeological and geochemical uncer- sulphide content of a waste rock is compared tainties in this context. etc. initial steps Whitworth (2002).g. More recently. family.e. where and when will been abandoned but continue to pose a pollution mine waters decant to the surface environment or risk. One of the shortcomings to rebound predictions yielded by VSS-NET of GRAM is its inability to cope with (Younger & Adams 1999. 2002). Younger et al • automated identification of decant points 1995).g. modifications of the field applications. geochemical processes in and around mined systems. Gandy & Evans 2002). Full-scale application of kinetic geochemical Adams & Younger 2001. Younger et al 1995. In the popular in surface mining applications. (2001). this volume contribute to the struggles to diminish pally calcite). highly specific conditions. Add to this the usual absence of our experience in addressing the principal hydro- historic ABA results for the strata formerly geological question that arises from the cessation disturbed by deep mine workings that have long of dewatering of mines: i. and applied in (e. modelling to the derivation of management plans • (three-dimensional) 3-D visualization soft. Standard geochemical modelling • a priori identification by applied geologi- codes. the discharge points from any 2001). it cannot be regarded as being concept to incorporate this phenomenon anything more than a 'first approximation' have been developed by Banks (2001) and (Younger 2000b). physically based ground- have inhibited their widespread uptake.g. CHARACTERIZATION AND PREDICTION OF MINED GROUND 7 Sherwood 1997). Younger et al 1995. Younger & Adams 1999. which has so far been attempted in very of mine water flow paths and outflow zones.

wess.g. monitoring well completion. so that the resultant monitoring wells too often yield These problems have all been identified by data of little or no use.g. Surface mines. 2002). the importance of head-dependent inflows in governing the rate of flooding of workings after abandonment (for further discussion of this point. YOUNGER & N. Sherwood 1997). 2001. operating areas of deep mines. However. Dumpleton et al 2001). this is a serious source of uncertainty. Adams & Younger purposes of borehole drilling. suspicious that the borehole is a clandestine exploration borehole rather than a benign The question of when such decants will become monitoring well. operations are a key issue in relation to wood & Younger 1997. This is because mine owners hand. There is • They are rarely situated in the most optimal always a risk of over simplification on the one locations.8 P. ROBINS terrain models (e. have often had to fight Robins et al. Shortcomings in data needed to characterize see Younger & Adams 1999. we have found that it is often . Younger et al. although rarely dis. These may be grouped quarries are often drilled by the company's under three headings: blast-hole drillers during periods when • shortcomings in data needed to character. demands from regulators for extensive Banks 2001. Whitworth 2002) the parameteriza. pulsory access (although we are not aware There are three principal areas of hydrogeo. Whitworth 2002). Notwithstanding their drilling pro- • scarcity of information on hydraulic pro. tion of models for real systems remains fraught by effective monitoring may require that issues of system complexity and the general regulators exercise their powers of com- inadequacy of basic information.) Such limitations on the operative can only be addressed by some form of ability of minor operators to actually predictive hydrogeological modelling. (Opponents 2001) and/or digital geological structure of a surface mine development may well be models (e. The modeller therefore often harder for mining permits and will thus usually struggles to simplify the vast amounts of be better served with monitoring wells.g. With regard to the availability of data. their ordinary workload is temporarily ize the hydrogeology of mine systems. L. It is extremely rare to find extensive plans into digitally available. Younger & Adams 1999. Dumpleton et al. hydrogeologically relevant data suitable for modelling purposes groundwater monitoring networks outside the are substantial (e. Banks 2001. Although monitor the impacts of their dewatering suitable simulation codes now exist (e. with the assumptions about piezometric head distri- aim of identifying priorities for future data butions in and around mine workings. blast-hole drillers are rarely au fait perties and processes. S. of mined ground to determine optimum post. and only an experienced suitable locations outside of their curtilage. 2001. previous workers (e. Adams & Younger In the face of absent or poor piezometric data.g. monitoring networks. Sherwood & Younger 1997. being more conspicuous. 1995. points of potential hydraulic these monitoring wells: connection between otherwise separate volumes of mine workings. monitoring wells around closure management. Whitworth 2002).g. • To save money. the hydrogeologist is often faced with having to cussed in any detail. with the intricacies of hydrogeological • problems of scale. welcomes from aggrieved neighbours to In addition to these 'hard' sources of whom they make requests for access for information. 2002. practitioner is likely to strike the right balance and may not receive the most cordial of with any alacrity. Sher. Sherwood & Younger (1997) have noted that predictive At the other extreme. light. of any cases where this has occurred in the logical uncertainty surrounding any investigation UK). the hydrogeology of mine systems Johnson & Younger.g. a retrieval of all available hydrogeological modelling of mined systems mine plans for even a relatively small area of typically suffers from both a lack of certain types multiple-horizon mine workings will generally of information (principally hydrological data) yield a bewildering volume of paper. We will now consider each make unsubstantiated (and potentially drastic) of these sources of uncertainty in turn. information to the point of identifying a few there are a number of common problems with critical features (e. or the retention of unwieldy degrees of rarely have property rights in the most detail on the other. The and a super abundance of other information challenges associated with transforming these (mine plans). Given collection. in some cases.

ments will also have been keenly aware of the positions of aquifers or bodies of old flooded Scarcity of information on hydraulic workings in relation to active workings. • drainage of the affected strata into the we would rank the critical data needs for mine dewatered mine voids leaves the strata in water modelling in the following order of an unsaturated condition. of the predictions made to the quality of the data used. the enclosing strata. ing strata..NuttalUfa/.g.. ultimately proved never to have been excavated Robins et al 2002). Adams & Younger 2002). There is a fundamental paradox tational resources permit. A classic case of country rock which together constitute a mined misleading anecdotal information is presented hydrogeological system. by the time units. as • post-closure water level monitoring data. hydraulic conductivity and storativity of Sherwood 1997. payable where the working was • historical evidence of pre-mining piezo- rendered particularly uncomfortable by the metry (from shallow wells and springs). rence during and after working. These difficulties are by Adams & Younger (2002). However. these will have been the subject of repeated 'wet • local rainfall recharge and the ways in working claims' (i. but they all undertaken using conventional methods (e. presence of dripping or lying water).g. for instance. Davis & Battersby is always valuable to establish the sensitivity 1985.g. Burke & Younger properties of the mine voids and their surround- 2000. Younger et al. Where compu. voids. which this is transmitted to voids at depth. but which balance over annual time-steps (e. Sherwood & Younger 1997). using Monte summarized as follows: Carlo methods or some other stochastic simu. which may be modelling in probabilistic mode. and properties and processes may well have been involved in drilling boreholes to assess the water pressures in such With the best data-set in the world uncertainties workings. it dewatering system (e. not least because of the needs to be treated with caution. when the rising mine water reached them and • substantial over-or underprediction of the responded as it would to unworked crystalline rates of mine water rise in an abandoned bedrock of very low storativity (Adams & system (e. 2002). areas of goaf. From our experience with such modelling efforts. decreasing sensitivity: • it is not possible to measure the saturated hydraulic conductivites of unsaturated • the geological framework. this happens the original motivation for • mine geometry (defined by abandonment predictive modelling will have been over- plans). hydraulic relationship of mined ground to • while these can be measured once the adjacent country rock including aquifer strata are saturated once more. etc.g. Younger 2002). of the workings were wettest. especially difficulties inherent in assessing and conceptua- where the information is second hand ('My lizing flow processes in the complex of voids and Grandfather told me . requests for wage sup. • difficulties in predicting the required Given the variable sources of data used in the pumping rate for a new (or renewed) hydrogeological assessment of mined systems.'). and estimates of void space volume taken by events. who report a case manifest in problems such as: in which some supposed ancient workings were • water budgets that cannot be brought into widely believed by miners to exist. ultimately derive from uncertainties in hydraulic Sherwood & Younger 1997. it is probably best to inherent in the determination of the hydraulic face the uncertainties squarely and undertake the properties of mined ground. anecdotal information would still remain. Where numerical models have been The detailed reasons for these difficulties are constructed such sensitivity analyses can be multiple and vary from case to case. which are derived from In rare cases. boreholes drilled into still-saturated strata . Miners usually remember vividly which parts • operational mine dewatering records. plements.g. Dumpleton et al 2001).. 1995. and interconnectivity between void spaces. Officials • anecdotal evidence of mine water occur- who had responsibility for mine safety assess. and thus the voids. • we know that mining changes the saturated lation protocol (e.e. CHARACTERIZATION AND PREDICTION OF MINED GROUND 9 necessary and desirable to fill substantial gaps • sources and piezometric characteristics of in data availability for areas with a long mining lateral head-dependent inflows into mine history by resorting to anecdotal information. hydraulic tests undertaken in these.

ROBINS overlying active mine workings have revealed relating mined void volumes to average annual the temporal variations in transmissivity (and. to develop physically based models in which • The inherent difficulties in assigning individual mine roadways/shafts are resolved appropriate 'active surface geometries' and represented by one or more elements of a to those fractions of the rock mass (such pipe network (e.g. (e. Younger & Adams are hysteretic. Booth 2002. At Evans 2002) and simple reacting fraction the coarsest of scales. Wolkersdorfer 2002) and by changes in natural water quality (Nuttall el al.g. The difficulties in applying standard geo- Problems of scale chemical modelling packages to real mine water cases has already been mentioned. Loredo et al 2002). Saaltink et al To overcome these problems it is possible to 2002).e. each of which has been of potassium jarosite (which help to buffer deliberately developed for application at a given pH around 1. much more slowly than precipitation (e.g. the system. very simple models representations for pit lake wall rocks (which may be no more than manual calculations have also proved to have some predictive . 'ponds' of highly intercon. Banwart et al interpolation of values for unmeasured points in 2002. Such efforts have revealed the very short distances in mined ground.g. Dumpleton will inevitably necessitate frequent updating of 2002). comminuted waste rocks (e. Very real persistence of the following problems. Whichever by deformation of the rock mass due to void approach is adopted. Adams & Younger 2001).g. from a single dewatering well to kinetics of several key sink–source reac- an entire inter connected coalfield.g. Banks 2001). Northumberland. S. Nuttall et al 2002). with dissolution occurring 1999. they are highly unlikely to be available at waste using simple mass-balance calculations sufficient spatial density to support meaningful (e.g. Adams & Younger 2001. At the other extreme.g. Robins tions for the major mine water pollutant et al 2002). At the very finest scales. Sherwood & Younger 1997. These diffi- Even where direct measurements exist from culties have prompted two extreme responses: the which to characterize mine void flow regimes first is to eschew superfluous detail and model and the hydraulic properties of the adjoining the release of pollutants from weathered mine strata. Hattingh et al 2002) and in the because extreme variations in hydraulic con- aquifers infiltrated by such leachates (Saaltink ductivity and storativity typically occur over et al 2002). recent results suggest apply the various modelling tools summarized that the precipitation-dissolution kinetics earlier in this paper. 2002.g. At as pyrite) that act as pollutant sources intermediate scales. While this problem is by no means attempts have been made to numerically model unique to mined groundwater systems. Younger et al 1995). species (e. Adams & Younger 2001. and can also be revealed by means of tracer tests Geochemical uncertainties and challenges between shafts (e.10 P. it is the generation and multispecies transport of acid especially vexing in this context both because mine waters and spoil leachates through both the measurement points tend to be somewhat more source bodies of mine waste rock (Gandy & sparse than in public supply aquifers and Evans 2002.g. For all of the uncertainties that beset assessments 2002). the customary lack of data closure/goafing (e. Sherwood 1997. Whitworth 2002). which problems. Standard nected workings can be identified and modelled 'shrinking core' models have been found using semi-distributed water balance models to produce adequate results for highly (e. Gandy & Robins et al 2002. Bowell 2002. storativity).8 in very acidic mine waters) scale (e. to recharge rates) may provide as good information a lesser degree. flooded voids can also be assessed using various A good example of repeated updating of rebound test-pumping approaches. For instance. L. The predominance of timing predictions is provided for the case of turbulent flow in flooded voids can be inferred Whittle Colliery. The hydraulic properties of abandoned rebound predictions in the light of the latest data. e. of mine water quantities. Banks 2001. therefore. Banwart & Malmstrom 2001. the position in relation to prediction of mine water quality is many times worse. by Parker by numerical simulation of the time-drawdown (2000).g. which are induced as any more sophisticated approach. YOUNGER & N. it may be necessary Saaltink et al 2002). exist in attempting to serve as barriers to further advances in predictive conceptualize and model the hydrogeological geochemical modelling: behaviour of mined ground over the full range of scales relevant to many management • Lack of information on the nature and decisions (i. data obtained from such tests (e.g.

computer technology can be confidently While the simple suggestions made in Table 2 expected to ease this problem over time. single runs can easily take propose will increase the cost burdens of mining 12–18h where six or seven species are companies. Table 2 summarizes some preliminary lines of those presented by Nuttall et al. We propose increase markedly in proportion to both that compliance with these regulations be the dimensionality of the transport domain augmented by the implementation of relatively (i. worthy significant progress in this regard is of respect from the very start of a mining unlikely to be made before a wider project. and of the mine life cycle. viewed as idealistic when one considers the tic. and Hattingh et al 2002) and rationale. polluted waters draining from abandoned mine which was co-sponsored by the Applied Mineralogy voids. Keyworth. 1985). entire bodies of regu- the transport of a large number of species lations have long governed mining below water is simulated using a multispecies geo. and includes a will be averted in future when these workings number of additional papers on related themes. As we conceive cannot truly be stated for flooded or it. which we term many other mining companies. Our are finally abandoned? We would argue that the own insights represent the fruit of much research various experiences gained recently in relation funded by NERC. However. Hat. long been practised for health and safety • Numerical limitations. that was held at workings be designed so that such problems the headquarters of the British Geological Survey. acknowledgement that closure is the longest part tingh et al. if a 1-D flow field is replaced with a simple additional measures. they will substitute a little simulated for more than 20 time-steps in a extra expenditure during the exploratory and 2-D flow field. Orchard 1975) and the safe design chemical modelling code. 2002. on 15 February 2001.g. analysed using inverse geochemical mod. ultimate post-closure phase would be manifest in ical data-sets have been acquired for different ways at different stages in the mine life representative mine systems (along the cycle. suggestions in this regard.e.g. it are highly preliminary in nature. bodies (e. but we hope pollution exposure risk assessments will be able the readers will find many signposts in the pages to advance beyond the current state-of-the-art as that follow. previously been subject to intense historic/ such as those which dominate the evol. the same 'defensive mine planning'.g. An appropriate level of respect for the range of comprehensive hydrogeochem. rather. therefore. multi species 3-D transport modelling considerable inherited legacies that beset the can be applied routinely for geochemical opening of new mines in areas which have systems characterized by rapid reactions. the to abandoned workings provide the basis for a Environment Agency. working phases of the mine life cycle for the analysis of uncertainty by means of substantially greater expenditures that would Monte Carlo analysis is generally out of otherwise be incurred after closure (when cash the question. With such slow run times. we think that the general approach of Banwart et al 2002). and is. and may be may well be several years before stochas. Run times and monitoring of mine waste tips. it will be many more years before to fully realize this potential. Even with closure environmental management problems. which arise when purposes: for instance. Davis & Buttersby ution of mine water chemistry (e. in order to minimize post- pollutant species considered. . Scottish Coal and new paradigm in mine design. It is beyond the scope of this volume limitations. presented by Holton et al (2002). EPSRC. along the lines 3-D flow field) and to the number of proposed in Table 2. the Coal Authority. the question must be asked: can new mine Group of the Mineralogical Society. While further advances in flows are generally at their most negative). basing environmentally-defensive mine plan- ning on hydrogeological and geochemical Given that predictive geochemical modelling principles has considerable potential for the is still subject to such considerable practical future. prehistoric mining (e. CHARACTERIZATION AND PREDICTION OF MINED GROUND 11 power (Bowell 2002). defensive mine planning starts from the partially flooded deep mines (e. Of course 'defensive mine planning' has elling techniques. the very latest versions of popular reactive We do not envisage that the measures we transport codes. Towards defensive mine planning This volume arises from a meeting of the Hydro- Given the myriad problems associated with geological Group of the Geological Society of London.g. UK Coal. Nuttall et al 2002). together with their 2002.

N> Table 2. Pre-empts possible future water quality liabilities.) . etc. or (ii) (a) minimization of long-term water make. (ii) ensure adequate after-use of exploration boreholes. easy blocking of potential post-abandonment hydraulic pathways which should help to limit rock-water interaction residence times and therefore keep salinities as low as possible Site preparation (i) Construct mine access features consistent with detailed design. (i) Allows minimization of long-term water quality liabilities by careful handling. alkalis. (ii) local 'over-dosing' with calcite stone dust and/or topical grouting of high. (ii) minimizes mobilization of acidic ions in mine water after mine is S-high-K zones to minimize later pollutant mobilization flooded Mine waste Make provisions for selective handling/careful disposal of the most pollution. which would management generating waste rock (using methods such as co-mingling with reductants/ likely be very long term in nature. will also make eventual decommissioning less emergency situations expensive to achieve Main phase of (i) Careful design of panels/pillars/benches to minimize the inducement of (i) Minimizes water make and all associated costs. by means of water covers/dry covers. either by: (a) efficient back-filling and sealing. of liabilities associated with site drainage. 2002) allow full assessment and planning of mitigation measures for any water management problems arising from dewatering and/or mine abandonment Detailed design Plan pillar locations and geometry of major mine access features to facilitate Minimization of deep circulation of waters after mine closure. Examples of 'defensive mine planning' measures (based on hydro geological and geochemical principles) applicable at different stages in the life cycle of a mine Stage in mine Proposed measures Relevance to long-term environmental performance of mined life cycle system Exploration (i) Assay the overburden for long-term pollutant release potential.and syn-mining hydrogeological data to et al. and. O2 exclusion. extraction excess water inflow from surrounding strata. therefore. (i) Minimizes post-closure costs to achieve management objectives. (ii) locate mineral processing and tailings/waste rock storage facilities in those (ii) achieves compliance with standard water quality protection portions of the site least likely to give rise to environmental pollution in policies of regulators. (b) by equipping them for hydrogeological monitoring purposes (see Kuma (ii) (b) acquisition of pre.

abandonment (ii) seal major mine access features at or just below anticipated climax water (ii) minimize deep circulation (and therefore salinization) of mine table position. long-term site management/after-use. (iii) consider the installation of (?replaceable) in situ reactive media in main (iii) maximize the potential for emergence of good quality water at shafts/declines to provide treatment of polluted drainage prior to surface the ground surface. (iv) make sure facilities are in place for monitoring of rebound and climax water (iv) secures long-term monitoring to allow early identification of any table positions problems/demonstration of system stability to third parties Restoration (i) Ensure that hydrological issues are given suitable prominence in restoration (i) Minimization of long-term pollutant release through restricting plans for underground voids and mine waste depositories. discharge. (i) Achieve stable post-closure water management system. (iii) implement any short-term intensive water treatment measures during the (iii) avoids any legal problems during the period of most elevated 'first flush' (see Younger 2000&. (ii) implement long-term (?passive) water treatment measures as appropriate (ii) ensure long-term attainment of water quality objectives in (see Younger 2000c) receiving watercourse . (ii) involve all relevant stakeholders in financial and institutional arrangements (ii) establishment of a secure socio-economic foundation for for post-closure site maintenance and monitoring activities. (i) Ensure long-term drainage routes are predictable and reliable. c) pollutant concentrations After-care (i) Implement post-closure site maintenance and monitoring activities. waters. access to acid-generating materials by O2 and/or H2O.Mine (i) Engineer any long-term preferred drainage routes for 'permanence'.

. 363-371.297–314. London. LOGSDON. (eds) Mine model of rebound in South Crofty tin mine. C. approximation using the GRAM model. & 198. Mansfield. & HAMILTON.G. The hydrogeochemical dynamics Geological Society. K. Laboratory and oxidation waters in central and eastern England. BOOTH.E. Geological Society. N. S. PL. M. Geological FERMOR. PRETORIUS. 157-174. London.14 P. Special Publications. J.O. 198. 2002. 198. 2002. Mine water rebound in South BANKS. R. PL. Reviews in Economic Geology. Nottinghamshie.. 33. ROBINS.S. M. L. 1992. 1991. 275-286. D. S. 1999. MERRIN. prediction and on overlying aquifers. 2001.S. & BANKS. Journal of Engineering Geology and Hydrogeol- BANKS. Transactions of the Institution of Littleton.A. & ROBINS. Shire. & EVANS. In: sequent flooding of Wheal Jane Mine in Cornwall. PL. Special Publications. (eds) Mine Water UK. London. & gical and geochemical consequences of the Cox. M. PL. the Total Environment. 94. S. Predicting mineral weathering rates at field scale The mine tailing accident in Aznalcollar. 198. KRANTZ. PL. 198. (eds) Mine Water YOUNGER.D. N.J. ogy. Environmental Publications. (eds) Mine Water Hydrogeology and Geochemistry.R. 73-97. Mine Water Hydrogeology and Geochemistry. M. 198. G. In: YOUNGER. Special Publications. M. Ground Water. 93-99. A strategy for DUMPLETON. p. ADAMS. 74. G. GRIMALT.. 2002. & MACPHERSON. N. & MALMSTROM. (eds) Mine Water Hydrogeology GRIMSHAW. N. supplies to the United Kingdom iron and steel chemical modelling for preliminary assessment of industry. 3–11. (eds) cal Society. London. 32.. South and Geochemistry. Special Publications. A135–A147. R.S. Mine Water Hydrogeology and Geochemistry. Opencast and Geochemistry. A75–A85. PULLES. Publications. Paradise BOWEN. R. London. 2000. C. D. Hydrogeolo- BROWN. N. 55-68..S. PL. 1997.S. Assessment.L.L. 2002. D. 89-97. Special the good. CO. London. alization and predictive modelling. A variable-volume. Geological Hydrogeology and Geochemistry. S. Coalmining in Britain. & ROBINS.S. London. & CROXFORD. K.E.C. S. Society. chemistry. Reopening of Volume 6A. North Pennines. Deposits. & BANKS. HATTINGH.. 2001. JOHNSON. the Slovak Republic: Novoveska Huta. In: MATHER. YOUNGER & N. S.D. Geologi- Cornwall. 198. W. N. N. 4. p.. head-dependent Nottinghamshire: risk evaluation using 3-D visu- mine water filling model. DUSSEK. Geology. N. Part A: Processes. RL. 128. 2002. S. R. M. lost? Assessment of liabilities at a uranium mine in Pollution resulting from abandonment and sub. DUMPLETON. The historical use of mine-drainage and pyrite. & ROBINS. BOWELL. Hydrogeology Journal. Prehistoric Flint Mines.. Sunshine Miners. G. (eds) Mine Water Hydrogeology and Geo. RL. D. BANWART.. K. & YOUNGER. 39. London.. (eds) Mine Water Hydrogeology quality: a case study . DUMPLETON. R. D.S. 159–185. & DUMPLETON. Exploration.A. 198.M.P. Special Publications. ARNESEN. RL. Science of for mine water risk assessment. & ROBINS. Africa. S. RJ.J. 75-88. & BAKER. In: YOUNGER. Journal of Geochemical and Metallurgy (Section A: Mining Technology). IVERSEN.A. S. EVANS.GANDY. 1996. PL.. Techniques and ogy. 307-319. UK. 1985. Ground Water. 34. The effects of longwall coal mining C. L. K. A physically based In: YOUNGER.R.S. In: YOUNGER. Effects of heavy metals on the rebound in the South Yorkshire coalfield: a first aquatic biota. London. YOUNGER. N.. Wheal Jane mine.A. 347-363. 2002. & SWART. 2001. 198. 137-157. (eds) Special Publications. In: YOUNGER. N. Quarterly 362-365.G. & YOUNGER. Transactions of the Institution of Mining minewater pollution. 17-45. Geological Society. J.L. 2001. In: Pb–Zn fluorspar mine. A. BURKE. & ROBINS.J. (eds) Mine Water Hydrogeology and E. YOUNGER. ADAMS.S. 39. Mine-water chemistry: Geochemistry. Special Publications. 149-160. D. R. In: PLUMLEE.. 1998. Water Hydrogeology and Geochemistry. YOUNGER.J. Special Publications. E. Special Publications. RL. London. of mine pit lakes. JONES. FERRER. Hydrogeology and Geochemistry..M. 2002. ROBINS. S. J. 2002. & YOUNGER.M..J.. 242. In: YOUNGER.L.' 198. PL. M. & ROBINS. 1999. & ROBINS.S.-T. PL. WALKER. LEIGHFIELD.W.Hlobane Colliery. & management of long-term post-closure water ROBINS.S. Geological Society. Iron ore BANWART. Society of Economic Geologists. numerical modelling studies of iron release from a United Kingdom.E.S. the bad and the ugly.G. 365-377. properties of the overlying Sherwood Sandstone. 1942-1992.HOLGATE. Geological Society. Quarterly (eds) The Environmental Geochemistry of Mineral Journal of Engineering Geology and Hydrogeol.. R. A. PL. 205-214. Special Princes Risborough. Geological Society. DAVIS. their migration. 110. & ROBINS. 2002. British Coal Opencast. 2002. N. Groundwater KELLY. Fingerprinting mine water in the abandonment of Frazer's Grove carbonate-hosted eastern sector of the South Wales Coalfield. . & BATTERSBY. P. P. Health Issues. GOLDRING. S. ROBINS References Mining and Metallurgy (Section A: Mining Industry). N. S. 249-261. (eds) Groundwater contaminants and Society. 2001. Geological Society. Effects of longwall mining in modelling ground water rebound in abandoned the Selby Coalfield on the piezometry and aquifer deep mine systems. & JUCKES. C. Geological Society. HOLTON. C.. & YOUNGER. BANKS. In: YOUNGER. KELLY. Hydro. spoil heap in County Durham. London. In: YOUNGER.

Special Publications. shire. The Power Game: The Struggle for Geochemistry of Fossil. In: Proceedings of the Con- Quality at Surface Coal Mines. Mining and Metallurgy. CHILTON. N. Ancient Mining. Institution of Acid Drainage Technology Initiative (ADTI).A. WOLKERSDORFER.. 315–326. Prediction of Water United Kingdom. Duckworth. Special Publications. & ROBINS. Geological Society. Coalfield closure and environmental consequence USA. Office for the Co-ordination of Special Publications. University of Newcastle. London. YOUNGER. Morgantown. B. Terence Dalton Publish- . Modelling Minewater Flow 327-336. R. R. ROBINS N. SHERWOOD. Wheat Jane YOUNGER.M. 99-105. Morgao catchment (Mieres. and Quality Changes After Coalfield Closure. 1994. J. 198. Modelling Depressurization of the north wall at the Escondida groundwater rebound after coalfield closure. Humanitarian Affairs. Virginia University. Aberdeen- London. R.S. 1975. J. London. BEALE. Modelling the oxidation of 198. In: YOUNGER. SMITH. J. Special Publications. (eds) Mine Water Hydrogeology Mission Report. KUMA. N. Institution of In: YOUNGER. N.. 1993. PL. Management. M.S. 3–6 September 1996. & ROBINS. 1995. Rotterdam. YOUNGER. P. 2002. 2002. & ROBINS. 198. Environmental Appraisal and Hydrogeology and Geochemistry. 107-119. FERRIER. Society for Environmental Geochem- ROBINS.D. Mine gas hazards in the surface Hydrogeological framewrok for assessing the environment. PL.S. J. UNEP. W. N. R. PL. Manual prepared ference in Minerals Metals and the Environment II. ELLIS. Water Quality Series No.. Geological Society. 1-29. HMSO. West ROBINS...L.47–60. N. adjoining mercury and coal mine spoil heaps in the Special Publications.the case in South Nottinghamshire. northwestern Spain). In: YOUNGER. .R (ed. Water Hydrogeology and Geochemistry. Geneva. et al. PL.S. 253-262. 6-8. In: YOUNGER. (eds) Groundwater in the Urban N. London. Report to the National Rivers Society.J. & PENDAS. Report of the National Rivers in the Leicestershire Coalfield. Cyanide spill at Baia Mare.. National Mine Land Reclamation Centre. B. Geological ORCHARD. AYORA. 198. (eds) Mine Mining and Metallurgy/Elsevier Applied Science. Unpublished PhD thesis. London. (eds) Mine Water 198. (eds) Mine Water Minewater Study. processes and Manage- chemistry. 198. & BOWELL. S. 1998. 337-346. Hydrogeology of Scotland. 187–204. 2000a. A.C. R. Iron. K. J.. MANSELL. (eds) Mine Water at the Deerplay (Lancashire) and Frances (Fife) Hydrogeology and Geochemistry. In: KNIGHT PIESOLD & PARTNERS LTD. Minewater pollution: the PARKER. A. United Nations Environment and Geochemistry. 109. 198. Hydrogeology and Geochemistry. SHERWOOD. SAALTINK. Publications. (eds) Society. & YOUNGER. DUMPLETON. 2002. South Western Region. PL. R.) 2000. I. PL. istry and Health.S. & YOUNGER. London. PL. Authority. K.mine water in surface and near surface J. sulphides in an unsaturated soil. Special Publications. POWELL. London. Coalfield abandonment: geo- and Metallurgy (Section A: Mining Technology). & YOUNGER PL. Geoscientist. DILS. J. J..S. 1993.. & Geological Society. Coal abandon. NEUMANN. R. Geological collieries. In: YOUNGER. A. C. West Virginia. Special ment. 165–170. 1990. 280-289.S. D. 1996. & ROBINS. Geologi. Authority. London. K. Programme. N. ment . 1994. C. The Mining Engineer. & WALKER. In: Geological Society. Chartered Institution of Water and Environmental London. 2000.. (eds) Mine Water Hydrogeology and Geo- Environment: Problems. Mine water tracing. 109. PL. PL. McGregor Science. Journal of the Authority. TAYLOR. Abandoned Mines and the Water Environment. cal Society. C. Society. NICHOLSON. ROBINSON. L. A.S. ADAMS. 61 -73. Technology). Chile. YOUNGER. R. P. CHARACTERIZATION AND PREDICTION OF MINED GROUND 15 KLEINMANN. Geological Society. (eds) Mine Water Hydrogeology Hydrogeological and geochemical interactions of and Geochemistry. 198.W. 2002. Resources. by members of the Prediction Workgroup of the Prague. JENKINS.S. Groundwater rebound NRA. PL.S. K. Special Publications. Retreatment Plant in Baia Mare. MIRA. PL. & ROBINS. CARRERA. Copper Mine. SHEPHERD. DOMENECH. 198. 170. 121-136. C. Spill of plume migration from a tailings dam in the West Liquid and Suspended Waste at the Aural S. Transactions of the Institution of possible environmental impacts of large-scale Mining and Metallurgy (Section A: Mining gold mines. In: ment of flooded deep mine voids by test pumping YOUNGER. Nuclear and Renewable Coal. C.. Balkema. PL. YOUNGER. J.J. The monitoring and model- Integrated hydraulic–hydrogeochemical assess. F. 2002. & ROBINS.J. London. 14. Mine Water Hydrogeology and Geochemistry. & ROBINS.. 2002.. Transactions of the Institution of Mining YOUNGER. (ed) Energy and the Environment. Exeter. VALDIVIA. 2002. DARCY. London. Geological Treatment Strategy. (eds) environment: some historical evidence from the Diffuse Pollution Impacts.A. ling of mine water recovery in UK coalfields. Republic of South Africa. Mine water . MCKELVEY. Structural influence on UNEP. NUTTALL. chemical processes and hydrochemical products. J. N. Working under bodies of water. PL. A228-A236.. (eds) Mine Water Hydrogeology and Geochemistry. 1996. 261–270.S. LOREDO. N. 4 (5). Assessment & ROBINS. & HITCHCOCK. S. 2002. & SAMI. PL. A219-A223. London. In: YOUNGER. London.S.J.M. Special Publications. 2000..A. WHITWORTH.the role of the Coal revenge of Old King Coal. G. & ROBINS. 10. N. N. 1997. & COLLS. Rand. Romania.B. ORDONEZ. 1997. 2002. HMSO.

J. R. ROBINS ers. Institution of Water and Environmental Manage- Journal of Contaminant Hydrogeology. Meinne': mine water hydrogeology in the Celtic YOUNGER.L.16 P. To pump and Environmental Management. Geological 1995.. 182. 9.R. K. Enviromental Agency. Predicting temporal changes in environmental management options for the aban- total iron concentrations in groundwater flowing doned Durham Coalfield. or not to pump: cost–benefit analysis of future YOUNGER. ment. PL. 1995.L.R. Environment Agency R&D Technical doned mines. & Science of the Total Environment. 2002. 47—69. A. (eds) Groundwater in the Celtic strata in aquatic pollutant release from abandoned regions: Studies in Hard-Rock and Quaternary mines in a major. Oordrecht. Bristol. YOUNGER & N. PL. London.S. Journal of the Chartered from abandoned deep mines: a first approximation. PL. Fife. PL.33. 95 — 104.J.H. 2000b. British Hydrological Society. R. oolitic. N. Mining and Metallurgy (Section A: Mining YOUNGER. 109. L. Geological Society. & JOHNSON. from Kernow (Cornwall) to Ceapp Breattain land: nature. for Chartered Institution of Water YOUNGER.405–415.L. N. Uisge Technology). burgh. Transactions of the Institution of Report W179. berthierine-chamosite.C. A. & ROBINS. Hydrogeology and Geochemistry.S. Pumping from the Frances and Michael Collieries. & LAPIERRE. Holistic remedial strategies for YOUNGER. & HARBOURNE. 2002. Pollution. BLACK. p. 309-326. Lavenham. & HEDIN.B. extent and preventative strategies. B. Edin- Kluwer. Hydrogeology. Mine water pollution in Scot. Symposium. (Cape Breton. R. S. 198. 44. BARBOUR. YOUNGER. M. 2001. A210–A218. 2. Predicting mine water short. RL. Proceedings of the Fifth National Hydrogeology Mine water: Hydrology. YOUNGER. p. p. lands. (eds) Mine Water YOUNGER. YOUNGER. London. p.A.L. The importance of pyritic roof MISSTEAR. siderite iron ore field. pL. & SHERWOOD. 4-7 September'1995. 2000c. In: Special Publications.S. & ADAMS. 251-267. S. Nova Scotia).D. " . 35-52. ROBINS.25–2.and long-term water pollution from aban. rebound. Special Publications. BANWART. (eds) YOUNGER.M. Remediation. Predicting the Consequences of Ceasing Society. 2000. 265. UK. Cleveland..

heads in a moderately transmissive sandstone declined due to mining but recovered fully afterwards. (eds) 2002. the general conceptual model but also shows creating inflow and drainage problems. Increased fracture permeability affects heads because upper aquifers in high relief areas lose water through fractured aquitards to lower aquifers. with variations. Northern Illinois University.niu. and because the higher permeabilities lower hydraulic gradients and up-gradient heads. probably because of oxidation and mobilization of in situ sulphides during the unconfined and recovery phases. may produce irregular. operates through two quite distinct mechanisms: cussed. heads in a poorly transmissive sandstone fell rapidly during subsidence and did not recover. This paper discusses the generalized conceptual Mechanisms of the hydrogeological effects model of the hydrogeological effects of longwall coal mining on overlying aquifers and the shallow of longwall mining groundwater system that has evolved from Longwall mining is an economic. but they are typically overlain by a zone of low permeability that prevents shallower aquifers from draining to the mine. Second. Effects in shallow aquifers not in direct contact with the mine can be simulated using readily available flow models. efficient form numerous case studies since the 1970s. water levels may recover due to closure of fractures and to recharge flowing back into the affected area. USA.S. London. shallow bedrock aquifers experience head changes caused by fracturing during subsidence. are low-pressure groundwater sinks during active term investigation at two sites in Illinois. At the other site. As an of underground coal mining that is characterized overview paper. Unconsolidated. The effects of longwall coal mining on overlying aquifers COLIN J. and some- the different responses that result from minor times depletion of aquifer resources. after mining has finished. The thods such as room-and-pillar (bord-and-pillar) general principles apply to other areas. localized subsidence long The paper is divided into two major sections. The main geographical focus is on the extensive subsidence of the overlying strata and Carboniferous (Pennsylvanian) coalfields of the ground surface. Mine Water Hydro geology and Geochemistry. After undermining. If the overlying strata are sufficiently described. and increase down- gradient discharge. is mining. tered. the mechanisms and impacts of the hydro. a long. pL. The heavily fractured strata immediately above the mine dewater. predictable. Increased permeability led to enhanced well yields. 0305-8719/027$ 15. Studies at two active longwall mines in Pennsylvanian coal measures in Illinois support the conceptual model. a secondary drawdown spreads out laterally through transmissive aquifers from the potentiometric low in the subsiding zone. In addition. hydrogeological responses varied at the site scale due to variations in bedrock–drift continuity. groundwater will drain into the mine. First. in which pillars of coal or re-stacked specific impacts may differ substantially because rock are left to support the mine roof. Predictions and monitoring schemes can be guided by the general conceptual model. The hydrological impact of longwall mining geological response to longwall mining are dis. it includes numerous references by almost total extraction of large areas of the and summarizes previously published studies by coal seam and by the resultant rapid. but mining. However. Aquifers in variations in the geological setting within the close contact with the mine are generally dewa- same coalfield. Special Publications. the author. DeKalb. & ROBINS. It contrasts with supported me- eastern and mid-western United States. Shallow aquifers separated from the mine From: YOUNGER. but which of geological differences. including information from previous drainage and subsidence. At one site. . but water quality deteriorated. BOOTH Department of Geology and Environmental Geosciences. 17 Geological Society.00 © The Geological Society of London 2002. IL 60115. New fracture void space takes up water. but must consider local hydrogeological variations. N. USA (e-mail: colin@geol. which illustrates the application of permeable. All underground mines studies in various coalfields. causing large head drops especially in confined aquifers. 17-45. 198. unconfined aquifers were not significantly affected by mining.edu) Abstract: The hydrogeological effects of longwall mines are vertically zoned.

age to it. deformation at different levels in the overburden New South Wales (Holla 1991) and Nova Scotia due to the characteristics of the mine geometry. a zone of low- extracts large areas of coal by using a deliberate. which distorts regional groundwater flow. tion of hydraulic gradients. and-pillar mining. Booth 1986. . confining characteristics and prevents hydraulic out area. Coe & Stowe mining. particularly if bear the load of the immediate roof strata. safe and economical than room. i. Bieniawski 1987). Younger 1998. which is well described in the mining 1984. Modern longwall mine (Singh & Kendorski 1981. typically sure arch onto the side. as behind them collapses into the mined-out void. These in turn lead to changes in potentiometric heads. there is con- UK and adopted in various forms in many other siderable variation in the nature of the stress and areas. overburden to the ground surface.g.18 C. 1979). Walker 1988. Within the overburden These movements are expressed in the defor.Matetic&Trevits 1991. 1994). several kilometres long by 150-300 m wide. These effects are due primarily the span of the openings such that the roof is to in situ hydraulic property changes within the supported between pillars of coal or re-stacked shallow bedrock aquifers. Rauch 1989). 1994). It became increasingly used in the USA through the 1970s and 1980s. Kadnuck 1994). joint apertures. Above that Longwall mining completely removes large level. changing the hydraulic properties of The strata movement is manifested as fractur- the strata and. deep mines in low. The ground at reported for some mines in Illinois (Cartwright & the advancing front of the trough undergoes a Hunt 1983). In contrast. J. observed over the last decade in coalfields in The strata above the immediate roof settle onto England (e. hydrological effects 1992. roof strata only up to a certain height. and as short-wall mining the equipment and the lithology. caused by the damaging collapse of the roof strata that the sudden increase in fracture porosity. fracture porosity and permeability.g. nant form of deformation over the longwall panel Longwall mining has been widely used in the is vertical settlement. The supports are advanced as the water levels after long-term depression can create coal face is removed. sequence of horizontal tension and compres- Longwall mining also affects the groundwater sion described as a 'subsidence travelling wave' system through subsidence and strata movements. under weak sandstone in Western Australia (Nik. Thus.e. the hydraulic ing. causing substantial changes in independently of drainage to the mine itself. and altera- extracted material formerly supported. even if the mine itself is dry. heads and groundwater flow patterns. consequently. For example. the overburden load is distributed in a pres- rectangular areas (panels) of coal. new ones. The roof over the narrow strip of working area After mining has ceased a mine may flood.g. On the other hand. longwall mining completely the mine. caused by increased and-pillar mines. this sequence occurs as volumetric dilation mation of existing fractures and the creation of followed by compression. for example South Africa (Hodgson 1985). across the width of the panel is temporarily producing a high permeability pathway that supported using moveable hydraulic jacks. and so on up through the 2000). Hill & Price 1983. not to drainage into rock. When entire mining removed to pillar-supported side tunnels on a areas are closed down. Booth & Spande 1992) have demonstrated A primary engineering problem of any under.g. In room. while keeping only a limited working connection between the shallow aquifers and the section of the mine supported. Burke & Younger the collapsed material. (Schmechel et al. The the mine is connected with others to form an coal is cut from the working face by a shearer and extensive flow network. Previous studies (e. there will engineering literature (e. ground- water flow patterns and well yields in shallow Description of longwall subsidence and its bedrock aquifers. as outlines the mined-out panel. Despite subsidence. BOOTH by confining layers may be unaffected by drain. is be a hydrological response in a shallow aquifer more efficient. permeability strata generally maintains overall controlled collapse of the roof over the mined. The domi- more recently in Utah (e. first in the Appalachian coalfield. preventing decline in potentiometric heads. where a permeability environments may remain dry for subsidence trough develops that approximately years after mining operations have ceased. strata. this is achieved by limiting permeabilities. front and rear abutments. However. the rebound of regional conveyor belt. bedding separation and changes in existing gradients. the hydraulic supports will bear the weight of razetal. (Reddish et al. and the unsupported roof substantial mine water discharge problems. a typical response of rapid (but often temporary) ground extraction is roof control. later in Illinois and Deformation zones above the mine.

The presence of the intermediate confining Higher strata sag but are essentially self-support. less. meability changes above a longwall mine in a drological response. was of the order of 90 m under valleys and 150 m 2. . Strata deformation zones above a longwall mine (Singh & Kendorski 1981). 1981). which corresponds to zones I fected. typically 2-8 times the Hutcheson et al 2000). theoretical setting of the Appalachian Plateau logical division of the overburden is into three type. although the critical freely than the deeper strata. in erally do not occur. several deformational zones have been was first noted (Fig. intensely fractured. directly connecting the mine to the surface gen- 40 times the height of the workings. typically devel. Singh & Kendorski little extensive fracturing. which However. a continuous deformation and overlying productive aquifers or surface zone which subsides coherently with water bodies (Holla 1991. Tieman & Rauch 1987. UK: 'goaf'). which can move more mine and shallow aquifers. structure and topo- The thicknesses of these zones vary with the graphy.g. Deformational styles are also distinctly Fig 1. highly per. subsides coherently with little fracturing. in which aqui. in which roof of longwall mines under lakes and the sea convin- material collapses directly into the mined. oped in a shale-dominated interval. Peng 1986. Van Roosendaal et al (1995) ability) characteristics and corresponds to observed that an overburden dominated by shales zone III above. lithology and from area to area. Time-related deformation at the edges of the termediate confining layer. under hilltops and plateaux. cingly demonstrates that a confining zone nor- out longwall area. value depends on lithology. a minimum thickness of the IV . which the strata break by vertical fractures Subsidence engineering guidelines dictate and horizontal bedding-plane separations.a heavily fractured zone. and glacial clay prevented the loss of water from 3.a caved zone. the effective confining separation can be much maintains overall confining (low perme. mally exists and that highly permeable fractures II . Coe & Stowe 1984. separating the wells with permanent 1.above that. minimum separations between longwall mining III . 60 m deep in Illinois. in their integrated model. relatively unfractured. For example.g. Similarly. Rauch 1989. in the appropriate geological setting. USA. The most important hydro. corre. typically 30. EFFECTS OF LONGWALL COAL MINING ON AQUIFERS 19 including the caved area (US: 'gob'. confining layer is necess- turing at the ground surface and in the ary to maintain hydraulic separation between the shallow strata. (1981) and has subsequently been affirmed by Rauch 1987): many researchers (e.a deep. . panel. shallow sand aquifers to a longwall mine only sponding to zone IV above. zone is a critical aspect of longwall hydrology. but provide a which simulated elastic strains and resulting per- convenient framework for understanding the hy. from which they are separated by the in. It ing.a near-surface fractured zone. fers are affected by in situ fracturing but not generally by drainage into the mine. water-level losses from those that were unaf- meable zone. recharged or only temporarily depressed. Thus.an intermediate zone.a zone of well-defined and open frac. I . 1) by Singh & Kendorski identified over longwall panels (e. and II above and dewaters into the mine. Booth 1986. Elsworth & Liu (1995) found that the zones: critical separation between the mine and well bottoms. The successful operation height of the workings. For example. . .

Whereas the inner trough Holla & Buizen (1991) conducted borehole undergoes both tension (dilation) and then com. ]Matetic & Trevits (1991) the opening) squared. geotechnical studies by the Illinois gradients. They should thus have a in the permeability values. the hydraulic effects. tension cracks open up. However. Els worth & Liu's which showed permeabilities increasing to tran- (1995) model simulated a zone of vertical separa. nections from the ground surface to the mine therefore. tion in the lower strata directly over the panel. and bedding separa. 2 x l O ~ 8 m s ~ 1 . packer permeability tests above a 424-m deep pression. include an approximately rectangular were unlikely to develop. the hydraulic conductivities at various intervals with a decrease of permeability down from the between 20 and 47 m above the seam increased in tensional peak. Thus. 9X 10~ 8 ms~'. decrease of 0.5 x ences should cause systematic spatial zonation of 10"8 a n d 5 . then declined to between 2. fractures at the ground surface. tially about 1 X 10~ 7 ms" 1 . The hydraulic conductivities were ini- the ground subsides rapidly. The strata in the interior were dominated pillars and ribs surrounding the mine. above the seam as the mine face approached. for example as measured by Whittaker et al (1979) in shallow (up to 25m deep) poorly permeable shale above a 54-m deep longwall Previous field studies of permeability changes mine in England. exhibit anomalous potentiometric ture dilation were separated by regions of lesser levels and increased vertical leakage. steps from about 2x 10" 9 ms~ 1 to peak at about Within the subsidence trough. inconsistent. ability (Booth et al. BOOTH zoned laterally around the subsidence trough. whose permeability (as a function of strain) in the abut. For in turn can create changes in the local hydraulic example. J. and an additional increase . stress differ. the behaviour can be The permeability of a single fracture is approxi. Changes in & Schwartz 1998). Neate (1980). These due to the first panel. bedrock strata are one to two orders of magnitude. Subsequently. but found no correla- greater residual increase in permeability than the tion between the two. the borehole. 70m and continued mainly as discrete. the outer margins (rear. Time-domain reflectometry (TDR) cable changes occur in the caved or heavily fractured studies distinguished shear and extensional move. increases in fracture the specific capacities of wells varied from a aperture or bedding-plane separation cause signi. ary peaks. 1998). Then. At Any point in the shallow strata above the panel the Lynemouth site (mine depth 207 m. then declined to about the strata. and fractures in the inter. respectively. maximum surface tensile and compressive Changes in hydraulic properties vary substan- strains at 22 and 60m. They ob- front) undergo only the tensional phase of the served increases in the number of fractures and subsidence wave. concluded that continuous vertical hydraulic con- A post-mining groundwater flow model would. studies were included in the review.8 times to an increase of 8. Australia. as shown by several studies in the mately proportional to the aperture (width of Appalachian coalfield. inter-aquifer leakage and the ground- State Geological Survey (ISGS) at the Jefferson water flow field. and may dilation or even compression. while the cantly affecting water inflow to the mine. sient peaks as the mine face approached. measured ment region and the near-surface zone at the permeability changes at two sites at intervals sides of the panel. Aston edges of the panel were largely characterized by & Singh (1983) and Aston et al (1983) reviewed subsurface shear fracturing and open extensional several studies conducted at English mines. Typical increases in permeability in shallow rounded by a border of more increased perme. Complex 1994). some of which may be per- County site in Illinois (see below) indicated manent. inner area of slightly increased permeability sur.20 C. declining to a slightly and zones of shear failure and increased vertical increased residual level. increased by about tions and increased permeability occur within an order of magnitude. test in- will first experience tensional stresses and exhibit tervals 15-55 m above the mine). At Wentworth (mine depth 54m). Regions of greater frac- inner zone. the location undergoes 2 X 10~7 m s~' after mining stopped 7 m short of compressional stresses. then. tempor- as the face undermines and passes the location. zone immediately above the extraction and in the ments. and the authors act as a corridor of preferential groundwater flow. inside the tially according to the strata position relative to edge of a 183-m wide panel (Mehnert et al the mine both vertically and laterally. signifi- by vertical extensional separations. At fluctuations began when the face reached within the ground surface. after the panel had passed.0 times ficant increases in fracture permeability. sides and final mine in New South Wales. permeability shear deformation as the face approaches. ior of the subsidence trough partially close up. and of a set of fractures conducted tests at two adjacent longwall panels to the aperture cubed (as described in Domenico 60–90m deep in southeastern Ohio.

Head drops due to direct drainage to the mine. shallow sandstone aquifers were probably tiometric low. hydraulically separated from the mine workings. (1986).8 to 845m2 day"1.4-0. indicates that inflow was responsive to shallow groundwater recharge events. as discussed below. levels to the sudden creation of new fracture void burden permits direct drainage and potential head space. 1986). Lancashire No. However. topo. which were probably due to drop in water level. although changes in permeability might Water journal. From Booth & Owili-Eger (1987) and Schulz (1988). The connections were probably under the major valleys where the Dynamic changes in hydraulic head mine was only about 90m deep. whereas wells with lower initial values had increases of between 2 and 4 times. with sudden drops in water level. be expected to relate systematically to the mining subsidence regime both spatially and temporally. varying from no change at one site to increases in transmissivity of 5 to 13 times at others. factors such as topographic relief and litholog. under the • direct drainage to the mine. there are certainly situations in 1983) have attributed the primary drop in water which a thinner and/or more permeable over. been explicitly explained. Pennsylvania (Booth The total increase in void space in a subsided 1984. Matetic & Trevits (1992) also studied two adjacent panels. and over most of the area the • transmitted drawdown around the poten. was in a valley bottom well 46m above the mine. from 1. Hill & Price confining layers. 225-258 m deep. Averaged over the whole overburden. and to • changes in hydraulic gradients due to adjacent longwall mining. as the relief at all sites is wells (Fig. declines. the increases being greater for valley wells than for hilltop wells. showed that averaged mine drainage the volume through which the ground has sub- quantities. Tests in several boreholes nested to depths of 46 and 91m showed that the valley bottom wells with the highest pre-mining specific capacities had decreases (by 0. but geological differences are important. In the both to greater rates of longwall mining and with Illinois studies reported later in this paper. Although shallower aquifers tend to be Head drops due to increases in fracture hydraulically isolated from the mine drain by porosity. with gradual. in the northern Appalachian area equals the volume of the extraction minus coalfield. but these can be attributed to causes other than direct drainage. There are several different the combination of the thinner overburden and a possible causes for this head drop: permeable fracture zone. The latter relationship clearly low. Reproduced with the permission of the Ground Thus. Correlation of mine drainage (MD) with shallow ed increased permeability or transmissivity over groundwater levels. 20 Mine.8 times) due to mining. seasonal water-level hydrographs of shallow graphy is not a control. mine in Cambria County. mining. adjacent plateaux and ridges the mine was up to • increased fracture porosity. Johnson (1992) found quite different results be- fore and after mining at four sites in Appalachia. study of the Lancashire No. in central Pennsylvania. slight increased permeabilities. 20 author's knowledge. However. Pennsyl- longwall panels in Appalachia include Johnson vania (BGL = below ground level). Various authors (e. million litres per day (Ml day ). 2. Other studies that have demonstrat. this . EFFECTS OF LONGWALL COAL MINING ON AQUIFERS 21 of up to 3 times from the second panel. Fig. corresponded ical variation complicate the relationship. At least one localized section of the mine underlying a valley The typical early potentiometric response of bed. to the For example. which were in the range 13-20 sided. had persistent inflow problems (Wahler & rock aquifers to undermining and subsidence is a Associates 1979). 240m deep. the relationship of this head loss. The greatest increase. • leakage from upper to lower aquifers Their water levels responded to longwall under- through fractured aquitards.g. 2). drop to the storage coefficient has not.

A truly uncon- extraction thickness of about 3 m at a depth of fined state presumably depends on whether air about 200m produces a ground subsidence of can enter the pore spaces. increase in bedding separation is a mere 0. Neglecting Moebs & Barton (1985) reported that water the minor changes on the magnitudes of the levels in 46 m deep wells in a shallow aquifer overall storage coefficient caused by the overlying a 230 m deep mine declined >30m increased fracture porosity. which would be pos- about 2m. Head drops due to the transmitted drawdown tures respond quickly compared to the less per. Furthermore. Thus. sand and aquifers. Most studies in the Appalachian actual drainage of pores) and storativities (which coalfield show that the head drops related to depend on the elastic compressibilities of water longwall mining are localized to within a few and the aquifer) the responses of unconfined and hundred metres of the mine panel. The only changes in water table in the drift of water into this new void space causes the aquifers were responses to subsidence-related hydraulic head to drop. the mine and well bottoms were at least 150m. if we which changes from a confined condition to a consider a hypothetical typical Illinois mine.g. resents a large percentage of the fracture porosity. which for a column of but sudden head drops just before mining oc- aquifer of unit area (A = 1) equals the change in curred in the poorly transmissive aquifer at the volume of (fracture) void space. the head simply relief 90–137m. This is only a small proportion of total (fracture plus intergranular matrix) porosity. This secondary effect expressed through the unit's storage coefficient spreads ahead of undermining and is therefore (S) . whereas in per unit area (A) of aquifer per unit drop (or gain) poorly transmissive units it occurs suddenly and in hydraulic head (A/z). The differences were clearly observed in our Illinois studies (see below). the head drops subsiding area is surrounded by an elongate occur quickly as void space increases. and the separations between drops very rapidly down to the top of the aquifer. Thus. changes in ground elevation (Booth el al. For example. The advancing potentiometric low in the meable intergranular voids. Because of the enormous differences in mag. 1998). . mining. the first response seen in any individual well. Our studies in Illinois showed hydraulically effective fractures and the low that water levels in unconsolidated. consider a unit area over the panels. the total overburden volume sible at least around a piezometer or well. where the aquifer with storativity 10–5 the theoretical drop overburden thickness was 200–335 m. The void. 'cone of depression' expanding as a 'drawdown' nitude of the head drop in each affected unit is effect through the aquifer. not in the shallow crease. BOOTH increase is unimpressive. Aquifers that are already much of the increased void space occurs in the unconfined are much less sensitive to this in- lower caved and fractured zone. Saline site. J. the potentiometric response to the that are consolidated (e. Sudden drainage did.1.the specific yield for an unconfined aquifer.e. and AVW is the rately transmissive aquifer at the Jefferson site. and an initial Thus. they are characterized by intergranular porosity nately large because of the low porosity of not by fractures.defined as head drop is transmitted further and occurs the volume of water released (or taken up) (A Vw) gradually in more transmissive units.01 m. whereas bedrock units at the same site tions and bedding separations. Tieman & Rauch (1987) plotted the extent of For an unconfined aquifer with specific yield 0.5%. unconfined storage coefficient of confined aquifers.22 C. Rearranging. but the frac. unconfined conditions. The mag. increases only an additional 0. the expected potentiometric effect of the overall porosity of (say) 10% (i.1 m. in which advance declines in head occurred early and gradually in the mode- where Ah is the change in head. the local in head is 1000m! In reality. change in volume of water. an threshold unconfined condition. Unconsolidated aquifers (e. dewatering of wells adjacent to longwall panels the drop in head is only 0. gravel) are also much less sensitive than those However. The new void space rep. closer to the site and time of undermining. effect. aquifers (sand units within the glacial drift) did space increase is not diffuse. Coalfield aquifers typically have relatively low nitude between specific yields (which reflect transmissivities. For example. 20m 3 initial increase in fracture porosity in a confined aquifer void space in a 200 m high column of unit area) is a rapid head drop to unconfined or threshold- would increase only to 10.g. The the storativity for a confined aquifer . but for a confined at a site in southwestern Pennsylvania. sandstones) because remaining void-space increase is disproportio. but is concentrated not respond significantly to longwall under- into a relatively small number of fracture dila. confined aquifers are very different. and not at column of a sandstone aquifer in which the all beyond a distance between 177 and 387m. slightly on the edge.5%.

gradient in the affected area must decline and/ or the specific discharge (throughput) must increase. state) feature. A similar phenomenon Within the inner subsidence trough. and concluded that wells are caused by increased permeabilities. due to mining 400m away. However. Heads will. but after mining (Tieman & Rauch 1987. system. Walker decline in upper aquifers that are not. Leavitt & Gibbens 1992. but down-gradient from the affected level aquifers. the overall effect will topography drives vertical flow. Fracturing of aquitards can cause water to drain from perched or upper-level aquifers down to lower-level aquifers or to local discharge Recovery of water levels after mining areas. paradoxi. dence-fractured areas.in effect. The most severe response (well dewatering) occurred directly over the panels. reported by Tieman & Rauch 1987). be seen within a local watershed as an increase in tion favours lateral flow and hill-side discharges stream and spring flow in the discharge areas (as (Stoner 1983. separate mechanisms coalfield have attributed head drops to this produce water-level recovery after subsidence: mechanism (e. compression and recharge. but the secondary Thus. based partial recovery occurs because the extensional on their study of a longwall mine 91–244 m deep (dilational) phase of subsidence is followed by a in an area of high relief (183 m) in Kentucky. compressional phase in which tension fractures The loss of water from upper to lower aquifers close back up. gradient. Matetic & Trevits 1990). Walker (1988) found to an individual well or site. to a water flow system that makes the generalized distance of about 300m. latter. squeezing the water back out below regional drainage level. caused by the dilational increase in fracture metric levels rose in some deeper piezometers porosity . The may increase. Studies show that from them down to the mine. Subsidence. However. It is this 1988. decline up- Head drops due to draining of perched or upper. the hydraulic 320 m. this sort of complex interaction between position aspect of recovery is dependent on connection to relative to the mine and position relative to sources of recharge and the ability of the aquifer . an early was suggested by Hutcheson et al. Thus. The water levels in the affected and found that the wells maintained their water area would be expected to recover as water levels because the water flowing into them from flowed back to the potentiometric low along the the shallow system exceeded the water draining temporary hydraulic gradients. stratification and the local ground- 'angle of dewatering influence' of about 42°. In another study in prediction of mining impacts so difficult to apply southwestern Pennsylvania. water levels may remain steady in lower recover at least partially within a few months aquifers that are draining directly to the mine. therefore. K) is linked to hydraulic mine (Booth 1986). Schmidt (1992) studied In addition. Werner & Hempel 1992). (2000). produced by the other factors. when the permeability is increased by (drawdown) effect was observed at a distance of subsidence-related fracturing. potentiometric levels and stream flows typically cally.g. The study by Hutcheson et al (2000) These closures partially reverse the head drops showed that. 20 hydraulic conductivity. and a related increases in vertical permeability shift the lowering of heads and loss of stream flow in the balance toward the topographically driven upland recharge areas. They do not affect head drops attributed to increased recharge in the subsi. Local groundwater flow systems area. the permeability (expressed as the In the author's study of the Lancashire No. EFFECTS OF LONGWALL COAL MINING ON AQUIFERS The upper dewatering zone extended out at an topography. which the authors of the fractures. angle of draw. Kipp & Dinger 1991). Settlement of the beds may also may in fact offset the effect of mining on the cause some closure of the bedding separations. Generally. heads may rise (as observed in valley wells in coalfields with high topographic relief reflect a by Johnson 1992) and groundwater discharges balance of conflicting controls (Booth 1988). but the stratifica. Changes generally unaffected unless they are within the in permeability per se do not alter heads. piezo. water levels in shallow wells gradient (/) and specific discharge (q) through were unaffected by supported undermining but Darcy's Law: declined in response to longwall mining. but not at 442 m. Several studies in the Appalachian At least two principal. the potentiometric depression due inflow and outflow zones in wells in a valley to the changes in porosity is a transient (unsteady above a room-and-pillar mine in Pennsylvania. that the first water-level fluctuations occurred when the advancing face was 120–180m distant (approximately equal to the thickness of the Head drops due to decreased hydraulic gradients overburden).

mines exploiting the Pittsburgh seam of the 6. decreased effect. in the Eastern Interior (2000). but no recovery occurred at compression which partially re-closes joints. most sites will exhibit permanent head losses active subsidence. gradients and altered groundwater flow paths. 64% returned to service without the need for 7.24 C. a rapid head drop during undermining. and the sudden increase somewhere. more or less sequentially: extensional variations in the permeability of an unconfined . horizontal missive aquifer. allowance for local controls such as topography. The field studies would apply to most areas. shear deformation recovery after the cessation of mining occurred at and opening of joints. and separation of bedding planes. The numbered sequence of obser- the intergranular pore spaces can contribute to vations can be considered either as separate long-term recovery. located in south- fractured zones. delayed drainage of water from sequence. a rapid partial rise in head due to fracture Appalachian Plateau coalfield. It corresponds to typical sites in mine. changes in successive times: permeability permanently affect the groundwater 1. and the Saline County confining layer. and found that re-closure. lateral pathways to sources of recharge. the potentiometric minimum. increases in permeability. valley wells were less affected site is recharged by water flowing back by mining than upland wells. a gradual recovery of water levels as the remedial action. Northern Illinois University (NIU) and the Illinois State Geological Survey (ISGS) have Sequence of Potentiometric Changes. the following composite model engineering geology over two active longwall (expressed in Fig. there is generally some 2. (1992) determined that recovery was more likely in valleys than on hilltops. Leavitt & Gibbens (1992) with the maximum dilational (maximum looked at 174 domestic water wells near longwall fracture opening) phase. These mines are referred to as the longwall mine. and is water supplies directly over total extraction abrupt in space and time because it repre- mining had any significant recovery. below which the aquitard had in Elinois. full cracks on the ground surface). especially along the ten. The only previous hydrological study of The mine is advancing towards the left in the longwall mining in Illinois was by Pauvlik & figure. sional margins of the subsidence trough (Booth 3. located in southeastern Illinois about 8 km Illinois. the pre-mining condition (which may flow system. USA been fractured./review 2001). a gradual decline in head in advance of et al 1998). A point in the shallow subsurface expe. comm. The changes in hydraulic heads reflect this In addition. with appropriate lasted 7 years. the caved and heavily the Jefferson County mine. 3) is adapted from Booth et al. The figure shows. first response to the transmitted secon- residual net increase. Werner & Hempel (1992) considered that recovery was likely in Hydrological studies of longwall mining wells below the regional water table but not in perched aquifers. This makes long-term recovery in fracture porosity. The change from stage appear inconsistent and hard to predict. the major confining layer and a central Illinois about 23 km southwest of the fractured shallow aquifer overlain by a shallow town of Mount Vernon. Although the compressional closure of fractures may partially reverse the dilational already be affected by earlier mining). coincident within a few months. case studies in which recovery typically occurred 5. Esling (1987). J. Similarly. from the bottom up. conducted joint studies in hydrology and izing the above. features at different places at an instant of time. and another site (Saline County) where the sandstone then continuing settlement of the strata which aquifer had very low transmissivity and restricted may partially re-close bedding-plane openings. At horizontal stresses (expressed as aligned tension our study sites in Illinois (Booth 1999). dary drawdown at the limit of influence. 4. Johnson into the aquifer. Summar. As a result of the various factors of mining due to the transmitted drawdown leakage through fractured aquitards. whereas sents a sudden spatial and temporal dis- Johnson & Owili-Eger (1987) reported several continuity in the aquifer properties. as noted by D. Coalfield. Banks (pers. rapid vertical subsidence a site (Jefferson County) with a moderately trans. only one of 19 the onset of tensional fracturing. mines in Illinois. but the potentiometric response sequence north of the town of Harrisburg. or as successive observations at the same point at Unlike the porosity changes. BOOTH to transmit water back into the affected area. In a 3 to stage 4 approximately coincides with study by Cifelli & Rauch (1986). who observed slight temporary riences. USA.

Groundwater flow systems are measure sequence of numerous alternating beds therefore sluggish. EFFECTS OF LONGWALL COAL MINING ON AQUIFERS 25 Fig. Conceptual model of potentiometric response to longwall mining. of shales. which is limestones. In addition. clays and No permanent change was observed. relief is low. Compared to the Appalachian coal- reasonable considering that the aquitard was field. From Booth et al (2000). the The Illinois Basin coalfield contains several Illinois coalfield is largely overlain by Pleistocene thick coal seams of Pennsylvanian (Upper glacial deposits. glacial–alluvial silt aquitard during subsidence. and the topographic Carboniferous) age. mostly tills. less by permeable sandstones. therefore. within a typical coal. siltstones. the stratigraphic sequence in Illinois is unconsolidated and. 3. dominated more by low-permeability shales and tive to changes in fracture properties. relatively insensi. and brackish–saline water is . Reproduced with permission of the Association of Engineering Geologists (Environmental and Engineering Geoscience}. sandstones.

Analyses are glacial deposits. 16 tests of various durations and rates minor thin limestones. hydraulic (Davis 1973. The test data were However. tested in the field for pH. Both mines worked the Herrin Saline County site the water levels did not (No. three panels studied. The bedrock is overlain by Pleistocene before and after mining. form aquifers • Pre. till lated by inflatable packers. 6) Coal of the Pennsylvanian Carbondale recover sufficiently to permit pumping Formation.26 C. The geo- is flat to gently rolling. with site. each about 183m wide by over 1530m residential and farm wells. coals and were conducted between 1988 and 1994 clays. Poten. et al. b). the overlying strata tests after mining. tial impacts of coal mining on groundwater • Pumping tests were conducted in wells supplies are therefore significant in a regional constructed over the centre lines of all sense as well as for individual users. • Approximately 500 samples of ground- Wisconsinan glacilacustrine deposits at the water were taken from the two sites at Saline site and a cover of loess. and shallow. Department of the Interior (1981) pro- wash sand-and-gravel units in the major valleys cedures were followed. Samples were still available for monitoring during the study. portant resource for rural residential and farm • In the slug tests. Boreholes were and buried valleys (Zuehls et al. horizontal ground strains and subsurface strains (using extensometers and TDR cables in cored boreholes). four long wall meters and wells drilled on site and of existing panels. but availability of equipment. BOOTH encountered at depths as shallow as 50-200 m geochemical analysis. with maximum local chemistry was discussed by Booth & relief about 15m. alkalinity. to identify hy- Because the aquifers are generally poor in steretic or non-linear behaviour that could yield and/or quality. with some pasture at the Jefferson pumping or bailer. were analysed by the Bouwer & Rice age annual rainfall around 100cm. sandstones. pressure or clogging. US productive aquifers are thick Pleistocene out. A sequence of tapped by large-diameter wells. followed for each interval. but they are too deep and mineralized to centre lines in NX (7. and the land use is mainly crop Bertsch (1999). Data area has a humid continental climate with aver. but methods varied County site. public water supplies have indicate permeability changes due to over- largely shifted to surface-water reservoirs. either still in use or abandoned. mostly Illinoian till with discussed below. but serious (1976. and mining of particular panels and continued for analysed in the laboratory for all major and several years afterwards. shallow groundwater is still an im. slug permeable sandstones and limestones of the and pumping tests: Lower Pennsylvanian and the Mississippian. The using automatic pressure transducers. analysed by the Hvorslev (1951) method. discontinuous minor sand and gravel units. revised Bouwer 1989) and Cooper droughts do occur (for example in 1988). (1967) methods. step-up and step-down pressures was poorly transmissive Pennsylvanian sandstones. but at the characteristics. and would surely become was near-instantaneously added to each more so if global climate changes were to create piezometer and the head response recorded frequent or extended drought conditions. Cartwright & Hunt 1983). and fresh logged. these are uncommon. The properties were determined using packer. At both sites. a small volume of water livestock supplies. The ISGS monitored several minor cations and anions. vertical subsidence (using surveyed monu- ments). 198 la. At the Jefferson County consist mainly of shales and siltstones. Sampling was by either fanning. Tests were conducted The two sites studied have broadly similar before mining at all the sites.At the Jefferson County site. In addition.6cm) diameter bore- be usable over most of the coalfield. core-drilled through the bedrock. J. Studies at each site were started before the temperature and specific conductivity. which underlie the coal measures. The hydrological studies Jefferson County site included potentiometric monitoring of piezo.and post-subsidence straddle-packer along the margins of the basin (Poole el al tests at all sites were conducted on the panel 1989). Most of the homes and farms located according to the nature of the well or around the sites have private wells into the drift piezometer sampled and the changing or bedrock. and sampling for long and separated by 61 m wide double-pillar . flushed clean and geophysically However. surged. The landscape various stages during the study. then water was injected through groundwater resources are generally limited to perforated pipe into vertical intervals iso- minor sand-and-gravel units within the drift. The most holes.

Previous mining by Panels 1 and 2 had already affected the potentiometric levels. However. No. 5. which is covered by 3–10m of glacial till. Trent et al. From Booth & Bertsch (1999). Our study began during the mining of Panel 3 and focused on the final panel. The mined coal seam at the study site was around 3 m thick at a depth of about 222 m. EFFECTS OF LONGWALL COAL MINING ON AQUIFERS 27 Fig. 4. of which about 2 m occurred within 2 months of mining. sand and gravel. 561-575) (expect for the US public-domain components). 4). Subsidence and strain behaviour were studied by ISGS (Mehnert et al 1994. Panel 4 began mining about 850m from the study instrumentation. Jefferson fractures in the marginal tension zones and County site. . barriers. particularly shear Fig. Map of the Jefferson County site. From Booth et al.1m. Ground subsidence at the panel centre line ultimately reached 2. The top of the Mt Carmel is at a depth of about 24 m below ground. and are dominated by poorly permeable shales. 5). were mined between 1987 and 1989 (Fig. which it undermined in February 1989. 1996). The overburden strata are relatively undeformed and flat lying. Reproduced with the permission of Springer-Verlag (Hydrogeological Journal 7. about 174m above the mine is the Mt Carmel Sandstone aquifer (Fig. which is 23–25m thick and is divided into a thin (3–5 m) upper bench and a thick (12– 16m) lower bench by a shale–siltstone unit. 4. Stratigraphic column at Panel 4. Subsidence was accompanied by considerable fracturing of the strata. and loess. and is overlain by a shale confining unit 15–18m thick. (1997).

in slug tests ST1 (October in the boreholes increased approximately one 1988) the pre-subsidence conductivity varied order of magnitude. several orders of magnitude due to bedding-plane Although different results from different separations that were observed in the drilling methods can be caused by factors such as low- behaviour and by geophysical logging.032 I s . In summary. whereas pumping tests sample shale) in a transverse line across the panel and a much larger volumes of aquifer. P411 in the tension zone) were drilled piezometers were screened in limited intervals in into the lower sandstone in 1992. and furthermore piezometer (P306) off panel.1 from the earlier to possible behaviour as an unconfined. layered.3281s.1 X from 1.3001s. Packer tests were conducted in several tests.1 and durations ranging from less than 1 values are greater than pre-subsidence values for up to 24 h (Booth et al. scale sampling effects within the a 15cm diameter test well (P350. For example.0 X 10-8ms-' before subsidence to were analysed by several methods to examine 3. bounded or double porosity of similar duration. from 4. and P305 into a deep the borehole. P304 screened Slug tests sample only the immediate vicinity of into the lower sandstone. Conductivity values phase. well sidence damage. The 3.8 m-' s' in P306 (off-panel) to 1 0 . It was concluded that the aquifer as a well increased from 0.073 I s . permeability skins in piezometers (Butler & Pre-mining boreholes drilled in 1988 included Healey 1998). but well P350 survived. The hydrological effects of the subsidence to the subsidence trough. Owing to subsidence.1 in pre- whole behaved as a confined single-porosity subsidence test PT5 (239 min at 0. have been reported by Booth & Spande (1992) Hydraulic conductivities determined from the and Booth et al (1997. 1998).m . 1998) The slug-test values for individual piezometers are the averages of Hydraulic tests. the panel became unusable because of sub.7 . whereas the region. 5 + m s . Despite the heterogeneous and sampling Sixteen pumping tests were conducted on well variations. J. BOOTH vertical bedding separation in the central trough by permeability discontinuity boundaries related area. at discharge rates varying from 0. and two P350 was open through the entire aquifer post-subsidence piezometers (P410 in the inner including the small confining unit. and the upper bench was influenced earlier.94Is. and the model calibration values centre-line boreholes T401 (pre-subsidence. leaky the later post-subsidence tests. subsequent tests. fractured). several piezometers (P302.6m s-' in P304 (inner panel region). The post-mining test results 10-8 and 2. The piezometers the differences in pumping rates and durations were used for slug tests and for observation meant that different volumes of the aquifer were during pump tests of P350.30 to defined. pumping tests using a single 0.1 m . were obtained by calibration of pump-test 213m deep) and T402 (drilled 6 months after simulations conducted using a MODFLOW subsidence 30m west of T401. slug and pumping tests are shown in Table 1 (Booth et al. P303. as well as in the packer tests noted aquifer. The piezometers over sampled by different pumping tests. the aquifer was unconfined.6 m s . The aquifer increased by approximately one order of post-subsidence test responses were influenced magnitude in the inner area of the subsidence . the specific capacity of the aquifer. (Booth & Spande 1992). completely is indicated by a spatial variation of approximately through the Mt Carmel Sandstone and at selected two orders of magnitude for the pre-subsidence intervals below that. below which the strata were extremely model for the bounded aquifer.1 . the permeability of the by leakage from the overlying shale. 1998). different vertical intervals were sampled.0 X 10. The holes were tested in 6 m intervals The natural lateral heterogeneity of the aquifer in T401 and 3m intervals in T402.1 ) to system.1 in post-subsidence test PT8 packer demonstrated that the two sandstone (258 min at 0. Similar increases benches were hydraulically separated and the are apparent in the slug-test results for compar- lower bench behaved as a separate confined able areas. to a depth of (McDonald & Harbaugh 1988) numerical 158m. Post-subsidence hydraulic conductivity 0.3 . In the 1988–1990 almost all comparable tests. Considering tests confined.7 X 10-7 to 9. open through heterogeneous aquifer were a more likely the entire sandstone on the Panel 4 centre line). the effects of subsidence were clearly P350.28 C.1 ).4X 10. but from 1990 obtained from recovery analysis of the draw- to 1995 the potentiometric level was above the downs in test well P350 varied between 2. Also. explanation in this case (Booth et al 1998).5 X 1 0 . However. and used in the lower bench. permeabilities of some individual horizons in Values at the same piezometers also differed the underlying shale and limestone increased by between the slug tests and pumping tests.0 X top of the sandstone. the conductivity values determined using the same geometric means of the sandstone permeabilities methods.

7 4.4E-8 Slug 2/89 _ 1.5 2.7 X l 0 .1 X 10 -5 PT15 3/94 Lower Sst 2.0 X10-' 5 1 .0 X 1 0 .6 1. Panel 4.0 X 10 .5 X 10 -5 2.2 X 10 -4 2.0 X l 0 . _ PT8 3/90 3.5 5.6 X 10.5 only PT16 3/94 3.0 X 10 -6 _ PT5 1/89 3.3 X 10-8 5.3 X 10 -6 3. 7 x l 0 .4 X 10 -6 1.2 X 10–6 1.4 X 10 -1 _ _ .8 X 10 -7 2.7 X 10 -4 3.9 X 10 -7 5. Jefferson County site Test Date Location relative to subsidence trough over Panel 4 (month/year) Centre line Interior area Tension zone Outside Pre-subsidence tests P350 P304 P303 P302 P306 Slug 10/88 _ 3. .4 X 10-7 3.7 X 10-7 3.6 _ PT4 10/88 1. 8 x 10 -5 2.0 X 10 -5 Numerical model calibration Inner area Tension zone Outside East –west 8. Hydraulic conductivity values (ms -1) from slug (ST) and pumping (PT) tests. 2.1 X 10 -6 1.6 X 10 -7 _ Post-subsidence tests PT7 3/90 3.0 X 10.8 X 10 -7 2.7 X 10 -4 and 2.4 X 10.4 X 10 . - P410 P411 P306 PT11 7/92 5.8 X 10 -5 2.1 X 10 -9 Post-subsidenc tests P350 P410 Transboundary from P411 P410 and P350 PT13 8/93 9.Table 1.7 9.9 X 10-8 1.4 a n d 3 .6 1.0 X 10 -5 PT14 10/93 9.1 X 10 -7 2.6 X 1 0 -7 3.7 X 10 -5 7.3 X 10 -7 _ .6 X 10 -8 North –south 6.0 X 10 -4 Slug 1992 .0 X 10 -5 1.5 X 10-6 72 X 10 -4 5.7 X 10 -7 .

approximately the top of the lower bench aquifer. . most probably due to adjacent and approaching mining. Jefferson County site. but had of Panel 3 (Fig. reached a minimum during the during active ground movements and a slight early tensional phase of subsidence and partially potentiometric adjustment to the new topography recovered after the face had passed (Mehnert et al. then slightly The potentiometric response to fracturing is recovered and then declined gradually as Panel 4 much greater in the confined bedrock than in the face approached. The by 4 years after the end of mining (Fig. 1997. 8). The differ- 1994). with the aquifer becoming about one order of magnitude to around 10 -3 . 6. BOOTH trough and two orders in the tension zone at the decline in head due to increased fracture porosity margin. 7). 4 m 1998). Potentiometric levels recovered Geochemical changes at the Jefferson site. When the site was undermined unconfined. J. water from the drift wells was fresh (total The sandstone water level had thus demon. dissolved solids (TDS) less than 600 m g – 1 ) and strated all the various responses of the conceptual of mixed cation. Panel 4. very quickly to about 35 m during the compressive Geochemical changes were described by Booth phase. It declined to about 34 m with the whereas the water level in W5A was unaffected. sulphate-dominant type. unconsolidated drift. (400m away). (1998) and Booth & Bertsch (1999). The water level in W5B fell probably already been affected by Panel 2 rapidly with Panel 3 subsidence in July 1988. there was little response in the drift sandstone over Panel 3 had declined as the mine water-table wells except for brief fluctuations face approached. The water level was deep) and W5B (shale. A similar response was observed in the well ences between drift and bedrock responses are and piezometers over Panel 4 (Booth et al. unconfined as a whole and threshold-unconfined for the lower bench. clearly shown by adjacent wells W5A (drift. From Booth et al. Water levels in the In contrast. Potentiometric changes. 19m) located on the edge initially about 19m below ground. the sandstone water level fell rapidly to about 43 m below ground.30 C. Potentiometric levels in the Mt Carmel Sandstone during mining. (1997). 6. Nitrate model discussed earlier: transmitted drawdown levels were generally high. as shown in Fig. nearest approach (152 m) of Panel 3. rapid because of contamination from agricultural Fig. created by the subsidence troughs. and full long-term recharge recovery. compressive partial recovery. Storativity values similarly increased during subsidence. then gradually to about 12 m below ground et al.

8. These were slightly bench of the sandstone. Long-term water-level response in well P350. Sulphate was almost the drift water due to mining. Illinois Basin (Graf et al 1966. a 60m deep sandstone well in an unmined through the whole aquifer. edge of Panel 3. bicarbonate levels remained approximately the mining) water is represented by samples from same. Water levels in drift well W5S and bedrock well W5D. During the post-mining recovery. 1990–2620mg I . .1 ) with an increase in sulphate water chemistry of the Mt Carmel Sandstone to 800–1272 m g l . but Piper trilinear diagram in Fig. P411. Two possible sources for brackish-fresh (900–1200 mg 1-1) and sodium the increase in sulphate were identified: leakage Fig. and about 200mg I-1 in P350. sulphate.1 (Fig. Sodium increased aquifer. The shale water in zero in R18. which was screened only in the lower mining samples from P350. Major ion proportions are shown in a from about 400 to about 600 m g l . some wells exhibited a slight reduction in salinity.1 . P350 became more brackish (TDS in the range Significant changes occurred in the ground. The changes were greater in P350. The water from the bicarbonate dominant. There was no apparent change in 1987. the Na-HCO3 facies is upper shale was a brackish (TDS less than 2000 to typical of shallow sandstones in much of the more than 4000 mg1–1). Jefferson County site. and pre. Poole et al 1989). open R18. From Booth (1999). Jefferson County site. From Booth & Spande (1992). Booth & Saric dominant type. the water in probably due to increased leakage from the drift. Reproduced with the permission of the Ground Water journal. mixed cation. 7. EFFECTS OF LONGWALL COAL MINING ON AQUIFERS 31 Fig. 10). than in piezometer area approximately 5 km east of the site. fertilizers and cattle feedlots. Panel 4. 9. The native (pre.

9. Reproduced with the permission of Springer-Verlag (Hydrogeological Journal 7. Fig. 10.32 C. 561–575) (expect for the US public-domain components). Piper trilinear diagram of major ion proportions in the Mt Carmel Sandstone. Carmel Sandstone well P350. J. From Booth & Bertsch (1999). Change in major ions in Mt. . Jefferson County site. From Booth & Bertsch (1999). BOOTH Fig. 561–575) (expect for the US public-domain components). Reproduced with the permission of Springer-Verlag (Hydrogeological Journal 7. Jefferson County site.

Reproduced with the permission of Springer-Verlag (Hydrogeological Journal 7. et al (1994. 11. six adjacent westward. so that the mine line subsidence was 1. and the depth of 122. Yields of all aquifers driven longwall panels of widths varying from are poor. deposits. capped by loess.0m and a roof about 35 m above the mined seam. limestones and several thin to discontinu. clays. The eventual centre- northwards at about 23 m k m . From Booth & Bertsch (1999). Panel 1 was 204 m wide with an Gimlet. and the mobilization. with inset showing instrumentation over Panel 5. Our studies concen. were mined successively north to south between Results of the hydrological studies at the 1989 and 1994 (Fig.1 . or At the Saline County site. Farms and homes around the site generally Saline County site have either shallow. 11). overlain by Wisconsinan glacial lake sandstone was unconfined. The bedrock strata dip gently mined in December 1989. and the maximum . of sulphate originating from sulphides Illinoian till containing minor sand and gravel oxidized during the period in which the units. Saline County site have been reported by Booth trated on Panels 1 (1989-1990) and 5 (1992. by water Panel 5 (Fig.5 m at the centre of the study site. 561–575) (expect for the US public-domain components) from the overlying shale into the upper bench of is about 122 m deep at Panel 1 and 97 m deep at the aquifer. large-diameter (l–2m). Map of the Saline County site. city water supplies. Trivoli. The principal sandstones are the (ISGS studies). typically about 3 m thick at a height of average mined-out height of 2. approximately 6m thick and 60–90 m The transverse line of monuments was under- above the seam. and lengths from 2286 to 3130m. drilled wells into sandstone. 2000) and of the geochemical 1993). The drift cover is 18-27 m thick at panel 1 and 12–18 m at Panel 5. water-table wells into the upper 6–8 m of drift. EFFECTS OF LONGWALL COAL MINING ON AQUIFERS 33 Fig.44m. The bedrock is covered by flowing back through the sandstone during unconsolidated deposits consisting mainly of recovery. and overlain by Pennsylvanian bedrock consist- ing of shales and siltstones with thinner coals. The mined coal seam was about 2m thick studies by Booth & Bertsch (1999). 12). 1997. and most homes are now connected to 188 to 287m. Subsidence and strata deformation at Panel 1 ous sandstones.

fluctuated 2-4 m during subsidence. Bedrock cores rose between 6 and 9 m. was shown by multiple bore. the strata were tight and would not accept water at valid injection pressures.e. including a 6 m interval in the Trivoli Sandstone where a hydraulic conductivity of 6 x l 0 . Subsidence damage re- stricted post-mining testing. TDR monitoring following mining. Thus.1 range. initially about 12m below ground.5 m. about 16 and 41 m mining. Figure 13 shows monitoring in the pre-mining borehole showed the response of one centre-line piezometer (BP3) that subsidence-induced differential shear and and its post-subsidence replacement BPPS. limited intakes were achieved in only a few intervals. residential well.' range. the difference between the amounts of depth of about 33. Permeability values were also obtained from pre. hydrological investigations at Panel 1 were Water levels in the sandstone piezometers were reported by Kelleher et al (1991).44 m. determined. Van Roosen. especially in the Trivoli Sandstone.34 C. US public-domain components). panel. Differential displace. but no further recovery showed an increase from 1.15m was differential displacement. Potentiometric responses at Panel I. Water levels in the deep (confined) drift piezometers extensional and compressional horizontal strains over Panel 1 were about 7 m below ground before were located. water levels then to maximize fracture interception. Stratigraphic column above Panels 1 and 5. However. respectively at angles of 10° to the vertical tension. shallow drift wells at the same distance had no hole extensometers to have occurred between discernible response to mining. Piezometers were installed into the lower drift sand aquifer (screened at depths between 19 and 22m). vertical intervals. J. i. the Trivoli Sandstone (various intervals between 42 and 60m) and one piezometer into the Gimlet Sandstone (81–84m). respectively. extensional displacement worked up through the The water level in a 42m deep bedrock strata from a first-break depth of about 53m. Slug tests and pumping- test observations in the sandstone piezometers before mining indicated hydraulic conductivities in the 10 7 ms . From Booth & Bertsch (1999). with a piezometer on the barrier pillar or the replace- change from almost no high-angle fractures to ment centre-line piezometer during 2 years of many. 1994) and Trent et al (1996).and post-subsidence packer injection tests conducted in the cored boreholes on the panel centre line. but the limited re- sults showed only minor increases within the same range of magnitude. of about 11 m in 1989 to a stable post-mining ment.1 was Fig. only about 0. Post-subsidence intakes were more Saline County site. of which 0. Before mining. However.10m was in the drift.64 fractures was observed in either the surviving bedrock per m from before to after subsidence. (1990. but the Trivoli Reproduced with the permission of Springer-Verlag (Hydrogeological Journal 7. had very little daal et al. recovering only about vertical subsidence between the top and bottom of 1. BOOTH different levels of the overburden strata. 561–575) (expect for the conductivities remained in the 1 0 8 m s . 12. declined rapidly from a pre-mining depth ing and bedding separation. advance response to the approach of the mine and Two boreholes were drilled on the centreline declined to depths of 49-55 m in the period from before and after mining to depths of 130 and just before undermining to the time of maximum 84m. The ISGS geotechnical and stabilized at between 8 and 11 m below ground.8 m s .5m in the winter of 1994–1995. Generally. Hydraulic tests at Panel 1. then inside the panel edge. sustained over more intervals.05 to 1. mainly during the 5 days around undermining. In some piezometers. out of a total subsidence of 1. located about 300m north of the Borehole geophysical logging indicated fractur. the overburden subsided largely as a single coherent mass. .

The response to longwall mining at Saline The ISGS geotechnical investigations were Panel 1 clearly differed from that at the Jefferson reported by P. was mined from May 1993 to July 1994 a sand and gravel within 6 m of the land surface. The sandstone was also was between –46 and +60 m of the line. The final break was in the drift at a depth the site at a local scale in both the sandstone and of 16. in some cases. expected to be more permeable initially. Saline County site. EFFECTS OF LONGWALL COAL MINING ON AQUIFERS 35 Fig. In this case. occurred during the period when the mine face mining was more abrupt. Final Panel 6. (1997).7- was nearest to the Panel 5 instrumentation in 1 0 . The two shallow separated to the south by a 40-m chain-pillar drift piezometers. Long-term recovery was. the drift units. which did not change significantly march 1994. starting in November 1992 and undermined on 2 and therefore the transmitted advance head drop January 1993. where the 15 and 41 m in from the edge of the panel.1 . compressional. located further south (up-dip). Kawamura in site. Carmel Sandstone. The tests of the piezometers over the panel were con- Trivoli Sandstone was less than 7 m thick and was ducted before. The actual responses did not occur as lithologies. From Booth et al. The Trivoli Sandstone at Saline was only a Booth et al. beginning when the face was 47m predicted. after mining. and its monitoring in 1995. the sandstone was wide and from 1 to over 100m long. and hydraulic conductivities in the range 10. after and. 2000). 13. Panel 5 was mined from April Sandstone were constructed across the panel in 1992 to April 1993 a distance of 3131 m on a face 1992 and undermined at New Year 1993. The horizontal ground own low transmissivity. Hydraulic tests at Panel 5. the Trivoli Sandstone is only about 20m deep and extensional strain being largely expressed as occurs at the bedrock surface in contact with the longitudinal tension cracks several centimetres glacial drift. which were screened in thin barrier. in Table 2 (Booth et al. The maximum transverse exten- A different response was expected at the Panel sional and compressional strains occurred about 5 site. A transverse line of subsi- quarter as thick and at least an order of magnitude dence monuments was monitored for 32 months less permeable than the Mt. strains were initially extensional and then therefore.6m several days after undermining. DeMaris and N. to be recharged more readily. to exhibit Subsurface strain monitoring using the bore- greater increases in fracture permeability with hole TDR cable indicated several small exten- subsidence. (1997). and to sional displacements in the shallow (<45m have a more gradual head drop and more rapid deep) bedrock at contacts between different recovery. negligible. Potentiometric response of Trivoli Sandstone to mining of Panel 1. had distance of 2606 m on a face width of 280 m. continued mining up-dip. The hydraulic results are summarized at about 20m below ground. J. Slug width of 287 m at a depth of about 97 m.37m on the centre line by the end of confining units. They also varied substantially across away. Ground subsidence primarily was much less but the head drop during under. and isolated from potential recharge sources by thick reached 1. One sandstone test Subsidence and strata deformation at Panel 5 well and 12 piezometers into the drift and Trivoli (ISGS studies).6 m s . The four deep drift piezometers . during encountered approximately at the bedrock surface subsidence.

2 X 10-9y -7. - Shallow drift P51S 7.0 X 10 7 -2.7x 10 8 P53D 12.3 X 10 6 _ _ .7 X 10 -8 -4.3-8. 0 X 10 .0X 10 9 _ 1.7x 10 9 2.8X 10 -8 -2.5x 10 .3 1.6x 10.9 X 10 l j -1.1 X 10 .0 X 10 8 -2.8 _ 7.4-23.5 3.9 1.4 X 10 .8 X 10.0X 10 7 _ 1.2 X 10 6 -3.8x 10 6 _ _ P5CD 7.2-13.5 X 10 7 -2.5 3.7 2.1 x 10 7 P54S 4.3 X 10.5 X 10 -8 _ 5. Hydraulic conductivities (ms 1) from Panel 5 slug tests.6 X 10 -9 8.9-23.9-12.3-8.6( 1/2/93) P52B 20.1 X 10 9 -1.3 X 10 8 5.8 7.2 X 10 8 -6.Table 2.4x 10.5 5.2 X 10 9 -7.6-18.1-20.6-5.8 1.7x 10 8 P54D 10.4x 10 P52D 14.7 -6.7 1.8 Deep drift P51D 15.5 1.8 l .7X 10 7 P53B 17.9x 10 7 _ 3. Saline County site Piezometer Screened interval (mBGL) Before (8/92– 1 2/92) During (12/92 –1/93) After subsidence (2/93-6/93 -1 Hydraulic conductivity (ms ') Trivoli Sandstone P51B 20.2-16.1 X 10 9 5.1 4.1 X 10 6 .8 1.8 1.3 X 10 9 P54B 15.0-15.3 X 10 -8 -3.8 1.

level varying from 2 to 6 m according to the local Their hydraulic conductivities did not change sig ground elevation.8 m s . but varied consider. The transverse section across Panel 5 and illustrates only measurable water intakes in GT4 were in a the relationship of the various potentiometric deeper fractured shale-limestone interval and in surfaces. sandstone. started to decline in December 1992 when the ably with position across the site due to litho.9-4. 9m in P54D) during the observed in the tension zone (P53B) or barrier tensional phase in early January. During active subsidence. In the southern In the Trivoli Sandstone. are a feature of apparent conductivity in the range 3 X 10~8-4 X low-permeability units that are unable to drain 10 ~8 m s ~l. located 46 m barrier areas (P53D and P54D) declined in inside the panel. Pre- the Trivoli Sandstone. The response of the deep drift differed sequently damaged by subsidence. In Trivoli Sandstone was flat. After mining. everywhere.6-cm diameter) bore. geological units and subsidence. and reflect localized pore-water com- to accept water. The water levels in the outer and 1(T8 m s ~ l . stone at 19. 2000) is a was indicated at depths of 32. in significantly between the clay till over the inner the post-mining replacement piezometer at this panel. the deep drift potentiometric level in the . The water became unconfined. a total of 17m over the varied from a clay till in the centre to a sandy till at panel and 12m over the barrier by the time of the southern edge. including a rise of about of the Panel 5 centre line. Only piezometer P52B.9 and 33. which had an apparent and post-mining conditions are represented by conductivity in the range 1 X 10 -4 X the potentiometric levels for November 1992 and 10~8 m s~J and erratic intake behaviour probably November 1994. GT4 1995).2m.5 m. In GT3. In GT4. located only 5 m away pression possibly due to shear deformation at the from GT3. water levels in the inner dle-packer tests in NX (7. and reached minima 6.8m lower (at leading edge of subsidence (Van Roosendaal et al 21m) and directly in the Trivoli Sandstone. depth. 14). EFFECTS OF LONGWALL COAL MINING ON AQUIFERS 37 were screened just above the bedrock surface. another surface of the Trivoli Sandstone (line b-b) is flat. sandstone became unconfined and showed zometer P54D) it was a sand-and-gravel unit with virtually no recovery except for slight (3m) conductivities in the range 10" 7 -10^ 6 ms" 1 . but most of the sequence was too tight quickly. at inner piezometer P51B. but the greatest head drop was over the centre line of the panel where the sandstone Potentiometric response at Panel 5.0m below ground. overlying the top of the Trivoli Sand.9 – 1 0 .8m. The pre-mining potentiometric related to the filling of discrete fractures. Over the southern barrier (pie. the sandy till in the outer zone and sand on site. The barrier (P54B) piezometers. The Trivoli Sandstone had an discussed by Booth et al (1999). The sandstone water levels nificantly with subsidence.8 X the barrier. pre-subsidence slug. which was sub. probably due to conductivity increased two orders of magnitude recharge from the overlying drift sand aquifer. the depth to water contrast. In contrast. mine face was — 300 m away. The anomalous piezometric rises. but the edge and barrier. limestone. with conductivities in the range undermining (Fig. the post- the shallow drift piezometers at the panel mining surface (line d'-d') had declined most at fluctuated considerably during undermining. the top of bedrock was 1. piezometers declined only slightly (less than 3 m) holes GTS (6 months before mining) and GT4 (7 in December. Water levels in sloped south before mining (line d-d). barrier piezometer (P54D) the water level test conductivities were mostly of the order of recovered quickly by about 4m in early 1993 1 0 ~ 8 m s ~ 1 . rises in the winter of 1994-1995. depressing the heads in the drift but The pre-mining potentiometric surface of the cushioning the head drops in the sandstone. piezometer head eventually stabilized at 7. 15m above their previous minima and over- the subdrift bedrock surface was in shale at a depth flowing at the ground surface immediately before of 19. Over the panel.7 X 10~ 7 ms~ 1 ). The potentiometric surface table in shallow drift wells 200-300m off the of the deep drift was 2-4m below ground and panels did not respond to mining.6m Permeabilities were also obtained from strad. the 1 0 . the conductivity dropped back to 7. However. The lower drift over the panel the mine approached. undermining. cored to 61 m. where the head relationships rapidly stabilized and returned to a normal indicate leakage from the sandy drift down to the seasonal range of 0. Open fracturing or bedding separation Figure 15 (from Booth et al. the and thereafter remained high. manifestation of the low porosity of the fractures The post-mining surface (line b'-b') was lowered and low permeability of the matrix.5 m. and fell rapidly as logical differences.1 . showed a permanent increase in response to mining concurrently with the bedrock permeability (one order of magnitude to potentiometric response. these interior water levels was stopped at 40 m at a high-angle fracture within declined gradually to a depth of about 4. then fluctuated erratically during months after mining) drilled vertically just south the undermining period. No significant changes were (12m deep in P53D.

Three years after at the Jefferson County site. circulation. sandstones progressively down-dip observed by ^OOmgl" 1 ). and after mining from off-panel private Conclusions from the Illinois studies wells in the Trivoli and Gimlet Sandstones (Booth The Jefferson County site. Saline County site. J. bicarbonate and the range 1200—1900 mg I" 1 . rather than any 1300-1900 rngF1 in a 43m deep well. sandstone water facies with different depths at this In the shallow piezometers over Panel 5 it was site corresponds well to the successive facies in slightly more mineralized (TDS 600. and in calcium. The water in the mining. almost identical to the zone at P53D. mixed cations and chloride. The deep drift wells over in nitrates. with relatively higher Water in the shallow drift wells off the panel was sulphate in the 20m deep piezometers than in a fresh (TDS <1000mgr 1 ) with mixed cations 67m deep residential well. Potentiometric response of Trivoli Sandstone to mining of Panel 5. The steep hydraulic gradient at the facies with TDS > 2800 nig F1 in a 91 m deep edge of the panel indicates that it is the low well. sulphate facies with TDS transmissivity of the sandstone. head drop in the deep drift was in the tension 35% Mg and 65% Na. 16) in a mining with only slight declines. Samples were obtained from the Trivoli piezometers over the panel before and after mining. the potentiometric levels had not Gimlet Sandstone ranges from a sodium chloride recovered. where the mine was & Bertsch 1999). there leakage from the overlying drift. the water in the to mining was a slight increase in bicarbonate Trivoli piezometers over the panel had increases relative to sulphate. The variation in and bicarbonate.or sulphate-dominated anions. Cation proportions from all approximately 220m deep. all suggesting increased downward sulphate. However. clay till over the central panel responded to sandstones plot on a Piper diagram (Fig. where much more water flowed back drift and moderate to high in the shallow drift. to a mixed cation. Nitrates chemical changes were less than at the Jefferson were commonly present at low levels in the deep County site. Reproduced with permission of the Association of Engineering Geologists (Environmental and Engineering Geoscience).38 C. The water in the Trivoli Sandstone was a marginally brackish (HOO-lSOOmgF 1 ) Geochemical results at the Saline County site. From Booth et al. (2000). which is influence of later mining to the south. exhibited hydraulic . The greatest straight line from 100% Na to a point of 40% Ca. (1989). magnesium and Panel 5 had slightly brackish water with TDS in sulphate relative to sodium. BOOTH Fig. geo- was no obvious change due to mining. 14. probably due to increased vertical cation evolution trend observed in the sandstone leakage through fractures. and the only apparent change due Poole et al. that is most more likely in a zone of active groundwater responsible for the lack of recovery. After mining. sodium bicarbonate type.to bicarbonate-dominant anions. in to the aquifer during the fuller recovery.

and post-mining potentiometric levels. . Section across Panel 5. From Booth et al. Reproduced with permission of the Association of Engineering Geologists (Environmental and Engineering Geoscience). showing pre. (2000).Fig. 15. Saline County site.

The primary poten. Piper trilinear diagram of major ion proportions in sandstone water. the secondary drawdown in advance of mining fined during mining and returned to confined was slight. panied by a deterioration in water quality in the sive Mt Carmel Sandstone increased substan. followed by solution magnitude along the tensional margins. cally and topographically to the Jefferson County ture space was opened up. However. and potentiometric behaviour in accordance with a physical enhancement of the aquifer. The subsidence at Panel 1 produced approached. Saline County site. and thus about an order of effect which spread outwards and was observed magnitude less transmissive than the sandstone at as a gradual decline in head as the mine face Jefferson. . then relatively shallow depth to the mine (122m). which may tially due to subsidence-related fracturing over have been partly due to leakage from the over- the longwall panel.40 C. site. form of increased TDS and sulphates. There was a The combination of increased hydraulic slight compressive recovery but minimal long- properties and recovery of water levels represent term recharge recovery. moderately transmis. During during the potentiometric recovery. the levels only slight increases in permeability. except that the sandstone units were thinner tiometric low generated a secondary drawdown and less permeable. and two orders of aquifer became unconfined. The increases were one order lying fractured shale but was probably mostly of magnitude in the interior of the subsidence due to oxidation of in situ sulphides when the trough over the panel. the active subsidence phase. BOOTH Fig. Reproduced with the permission of Springer-Verlag (Hydrogeological Journal 7. fully over a period of 4 years as recharge flowed The potentiometric response was modified by the laterally back through the aquifer. 16. Following undermining. with the conceptual model. 561-575) (expect for the US public-domain components). became uncon. this was accom- storativity of the shallow. but the heads dropped rapidly to the conditions during long-term recovery. J. The aquifer as lower transmissivity of the sandstone such that a whole was initially confined. potentiometric The Saline County site was similar geologi- levels in the sandstone fell rapidly as new frac. The permeability and improved well yields. top of aquifer with undermining. despite the recovered slightly in the compressive phase. From Booth & Bertsch (1999).

cularly pronounced in confined fractured aquifers • continuing loss of water to the mine. or. in because of the low magnitude of the storativity. How to mining was negligible. response and critically control the recovery. and subsidence. The initial permeability of the effect is still controlled by the magnitude of the Trivoli Sandstone was low. However. and less signifi. The potentiometric level in the sand. the mined panel. permeability on the flow systems are more subdued: there is less opportunity for the in- creased drainage from upper aquifers to lower Future research aquifers and discharge zones. transmissivity. not an effect transmitted up from the mine. such as tight con- prevented because intermediate levels of the fining units and lateral boundaries in the aquifer. through fractured aqui- secondary drawdown effect is transmitted tards to lower aquifers or down gradient to through the aquifer outwards from the potentio. which re- several different mechanisms. in situ fracturing caused by longwall mine • the location of continued mining oper- subsidence creates multiple. The critical factor appeared to be the The potentiometric drops are clearly greater over low transmissivity of the sandstone. but this is typically with sources of recharge. continuity and geometry of the hydrogeological units on scales as broad as upper-level aquifers to lower aquifers and the groundwater flow system and as local as the discharge areas through fractured aquitards. Essentially. EFFECTS OF LONGWALL COAL MINING ON AQUIFERS 41 At Panel 5. after the transient potentiometric depression caused by mining. years later. the subsidence strains or increases in per- oned the potentiometric impact in the sandstone. or nearly so. Variations in the hydrogeological pro- effect is amplified by the increased drainage of perties. levels can sometimes be lowered as a response • natural barriers to hydraulic continuity to direct drainage to the mine. interacting effects in ations. Our understanding of the hydrological processes cance to the effects of increased permeability on and effects of longwall mining has advanced . The sudden increase in fracture poro. Potentiometric stricts lateral recharge through the aquifer. However. this geology. However. A high-relief areas. but it can be delayed or prevented by several factors: General conclusions Longwall mining affects overlying aquifers by • low transmissivity of the aquifer. The potentiometric effects at the would be higher and that potentiometric recovery Illinois sites were focused on the areas over and after mining would be rapid because of easier immediately adjacent to the longwall panel. the response fracturing caused the confined bedrock aquifer of shallow. which can interrupt the physical or the aquifer. hydraulic continuity between the affected sity causes a transient head drop that is parti. but there was Recovery should be the normal state of affairs no long-term head loss. Increased fracture permeabilities cause decreased gradi. and in both coalfields the lateral related fracturing caused only localized minor limit of head drop is within a few hundred increases. This limit should not be directly con- stone declined rapidly with undermining. discharge zones. that do not adequately consider these local In low-relief Illinois. unconfined aquifers to become unconfined. aquifer and recharge sources. metric low in the subsidence zone. Otherwise. topography of the ground surface. the subsidence-related At both Jefferson & Saline sites. up-gradient head drops and down-gradient response to mining subsidence is the local hydro- increased discharge. the Trivoli Sandstone was at the system gradients and throughput that are more bedrock surface in contact with the overlying pronounced where the natural hydraulic gradi- drift. metres. Water tables fluctuated suddenly or gradually this condition developed briefly as they accommodated to the changing depended on the transmissivity of the aquifer. recharge. The most significant difference between the Generalized monitoring schemes and predictions Illinois and Appalachian coalfields is the relief. unconsolidated. the most important factor in the ents. and no trolled by the depth of mining as it is a response significant recovery over the panel was observed to an in situ head drop in the shallow aquifer through to the end of monitoring more than 2 itself. but in our studies these drops the existence of a sand unit along the edges of the did not obviously correlate with the magnitude of panel provided some recharge that locally cushi. meability. the behaviour was similar to Nevertheless. In high relief areas. It was expected that the permeabilities ents are steeper. the transmitted potentiometric that at Panel 1. the effects of increased characteristics are liable to be erroneous. overburden retain confining properties. which enhances the topographically driven local panel strongly affect the initial potentiometric flow system.

Madison. Adams & Younger 2001). Esling. fracture characteristics and rock mechanics. remains a problem. Roosendaal and J. D. increasingly complex theoretical situations. USA) are already and by the lack of adequate computer facilities widely used to simulate groundwater flow in (the need to reserve time and space on main frames complex. location. which are less well immediate application needs are likely at the documented than the short-term impacts. 20 study was conducted under the integrated process of the hydrological effects of supervision of R. Parizek and supported by the longwall mining. Pennsylvania State University Mineral Conservation Section and the Shell Companies Foundation.almost all types. Mehnert.Younger 2000.saturated regions above the mine is problematical. el al. However. trans. Bauer. moment to use the readily available commercial • ways in which numerical modelling can be variants of well-validated groundwater flow applied to individual sites to predict im. The research group at Newcastle University has followed a somewhat different • more field data and expanded conceptual approach to modelling groundwater rebound in models to account for differences in the abandoned underground mine systems. P. readily available as visual interfaces such as Booth 1984) were limited by the lack of empirical Visual MODFLOW (Waterloo Hydrologic.considered separately from the deeper units and water flow (groundwater hydrology). variably logical impact of longwall mining require the link. T. ges in groundwater chemistry on overlying Consulting engineers and hydrogeologists with aquifers after mining. Kim and timing of potentiometric changes and long. Van flow. three areas seem particularly the off-the-shelf engineering and hydrological in need of further work: working world.models such as MODFLOW (McDonald & pacts and explain hydrological responses. However. Southern Illinois Uni- cal models for the individual stages of the model versity at Carbondale. mechanisms and effects. A. Curtiss and L. C. by apply- hydrogeological settings of the new coal. such models should generally be mechanics and hydraulics) and the changes in acceptable for the shallow overlying aquifers hydraulic properties to the changes in ground. USA) and Graphic unavailability of ready-to-use. validated numeri. Kelleher. well-tested numerical models available tributors to the Illinois research were ISGS personnel for the individual components of groundwater R. These models are being applied to term recoveries on a site-specific basis. Con- versatile. Increasingly sophisticated coupled numerical lating the scientific understanding into direct models that link the mining. D. B. M. However. there is now a which the scientific research into the hydrological great deal of empirical field data and we have a effects of longwall mining can be applied to much better understanding of the hydrological practical engineering problems. develop such models about two decades ago (e. heterogeneous geological settings of made model development and calibration imprac. Miller. Using such Darcian. but there are sophisticated. Liu & Elsworth 1997. Spande. deformation.Groundwater (S. ing of the strata deformation to the mining (mining and generally quite inappropriate for the non- engineering and rock mechanics). Attempts to the mine inflow boundary. inadequate conceptual Ontario. Harbaugh 1988). The research was .. P.ing different forms of mass-balance models fields in which longwall mining is being appropriate to pipe conduits. Geological and inflow to the mine) have been developed over variation between sites makes it difficult to apply the past ten years by research groups such as those generalized concepts or generic formulae to at Penn State University (Bai & Elsworth 1991. modelling demands are widely available. GMS (BOSS International understanding of the hydrological processes. D. J. 1997). BOOTH greatly over the past two decades. engineering practice. larger-scale con- carried out. It is a truism that our knowledge is never although it will probably remain difficult to perfect and that more field data are always translate them from the academic research level to needed. Demaris.water flow in the intensely fractured. data from field studies. Canada). Pattee. so that specific predictions hydraulic property changes and groundwater of the effects of planned mining at individual mine flow (including in some cases variable saturation sites can be made.42 C. the changes in Darcian flow through the mine openings them- hydraulic properties to the deformation (rock selves. MODFLOW variants. B. Illinois. J. magnitude Elsworth & Liu 1995. porous media models to model mine inflow and ground- Truly integrated numerical models of the hydro. However. R. and NIU graduate and powerful personal computers capable of the assistants E. J. predict the exact occurrence. A. or Darcian porous media to • more research on the long-term recovery the different system components (Burke & of groundwater levels and associated chan. Wisconsin. nected mine ponds.g. There are still no 'off- the-shelf models available for the whole The Lancashire No. Inc. Their application is one means by tically time-consuming). Bertsch.

2. M. December. 2000. (ed. wells. IMWA. C. Ground Water. 149-160. DEMARIS. southern Illinois. 1984. 1998. In: FERNANDEZ-RUBIO. C. Proceedings of the 34th Annual determining hydraulic conductivity of unconfined meeting of Association of Engineering Geologists. approximation using the GRAM model. C. Hydrogeologic mining.407–418.P. The Bouwer and Rice slug test . & BAUER. Recovery of groundwater levels effects from selected underground coal mines in after longwall mining. C.J..) (ed...) Pro- Program of the Illinois Department of Energy & Natural ceedings of the Fifth International Mine Water Cong- Resources (1988–1992) and by the Office of Surface ress. WV. Ground Water.P. 1984. & SCHWARTZ. Denver. Subsidence Due to Underground mining. EFFECTS OF LONGWALL COAL MINING ON AQUIFERS 43 funded under the Illinois mine Subsidence Research County. USA. 2000. A Numerical Model of Groundwater Journal of Engineering Geology and Hydrogeol- Flow Associated with an Underground Coal mine ogy. Illinois. Ground I. 1999. update.. CURTISS. Modeling of BOUWER. OH. 1976. 12. 7.R. 305–313. & ELSWORTH.. Groundwater Research.M. S. Illinois. Potentiometric and Well Association. & VAN modelling ground water rebound in abandoned ROOSENDAAL. Illinois Basin coalfield. Evaluating the impact Saline County. August 1986. & HEALEY. & RAUCH. 6. International Journal of BOUWER. Water Resources Research. 19–34. Environmental and Engineering permeability of Coal measures strata associated Geology. associated with longwall mining. (ed.. DOMENICO. CURTISS. C. Appalachian Plateau. K. C. Groundwater BIENIAWSKI. Sevilla. 637 Grant/Coop.W. Environmental in a heterogeneous layered aquifer setting. 2nd edn.J. In: PENG. 1991. CIFELLI.H. Illi- nois. Morgantown.M. ween pumping test and slug test parameters: scale BOOTH. (4). H. 1983. International BOOTH. A slug test for subsidence and groundwater flow due to under- ground mining.J. College Station.J. BOOTH. Nottingham. & HUNT. National Water BOOTH. J. F. University of Kentucky. Anomalous increases in mine systems. those of the author. & BERTSCH. Morgan- Spain. 263-269. 34. Department of the Interior (1991–1995: Contract Aquifer Response to Longwall Mining. & YOUNGER. The Pennsylvania State University. Library Report. 1992. John Wiley. 1988. P. T. & YOUNGER. aquifer property changes above subsiding longwall COOPER. BURKE. R. H. R. Association Conference on The Impact of mining Bulletin.) Mine. 1999. 249–263. A. 507–515. 383-394. 1986.J.N. C.D. E. DJ. Saline New York. John Wiley. R. Water and Environment: Proceedings Proceedings of the Second Workshop on Surface of the I. 1994. 1998. 11 – 18. L. 2. 243-248. L. Interpretation of well and field data Illinois State Geological Survey. abandoned underground mines on ground water. Agree- any views expressed in this article. mine panels. Illinois'. PATTEE. 33. 249–261. B. J. 1989. potentiometric response to subsidence above active 1983. In: REDDISH.J.R. A.J. D. C. R. Colorado.D. instantaneous charge of water. PA. Z. TX 981–990. Positive and negative impacts of longwall mine subsidence on a sandstone References aquifer. COE.. S. 1998.. MILLER. SPANDE.N.J. BAI. 1983. 11. 348-359. A reappraisal of subsidence. & RICE. P.T. and Chemical Hydrogeology. Springfield. C.M. Dewatering BOOTH.C. Office of Surface constitute an endorsement by either DENR or OSM of Mining. J. 561-575.J. 2001. 3. SINGH. & STOWE. & WHITTAKER.M.A..D. 71–81. DJ.1.J. E. Aspects of Coal mining in Illinois: An Overview.C. & PAPADOPULOS. Physical BOOTH. Environmental Geology.D. PL.. Ground Water. 596-606. IMWA. Strata movement concepts and the effect or artifact? Ground Water. Environmental and Engineering investigations into strata permeability changes Geology. New York. Strata Control in Mineral rebound in the South Yorkshire Coalfield: a first Engineering'. DEMARIS.J.S.N. on Ground Water. R. 1986.. & BERSTCH. ADAMS. 1987. R. September 1999. geochemistry in shallow aquifers above longwall DAVIS.J. Relationship bet- Dissertation. 26.C. In: Proceedings of the of longwall coal mining on the hydrologic balance. 36. CARPENTER. EGN 90. BREDEHOEFT. Ground Water. J.M. 1987.A. C. aquifers with completely or partially penetrating Chicago. town. & SARIC.M.W.W. & SINGH. 30. H. 304-309. 4. Geology Notes. T. 1997. Groundwater response to longwall mining. piezometric levels in advance of longwall mining ASTON. C. R.A. Quality of near-surface waters in mines in Illinois.J.an Mine Water. & SPANDE.S. BAUER. 362-368. P. . R. R. The effects of University. north-central West Virginia.W. Columbus. 24.J.J.P. In: Proceedings of the National Water Well Sedimentology and Reclamation.A. 39. 1987. C. with longwall mining.D.J. Response of a finite diameter well to an Water. PL. Hydrogeology Journal. H. hydrogeological impact of underground coal CARTWRIGHT.J. Hydrology.C. & MILLER. Pennsylvania Ph. West Virginia BOOTH. Ground Water. which are solely ment GR196171. West Virginia. in the Appalachian Plateau. BUTLER. International Congress. Water Resources BOOTH. S. This support does not Department of the Interior. 27. S. BU 145. National Symposium on Mining. 423-428. 1973.D.T. 1999. US Agreement GR 196171). Quarterly BOOTH. 223–233. C. Site-specific variation in the ASTON. BOOTH. September 1994. C. & Journal of mine Water. J. C. A. 1–14. Mining Reclamation and Enforcement of the US BOOTH. Association of Engineering Geologists. 1967. J. The effect of test cavity geology on the in situ longwall mining. (4). 35–40. USA. CURTISS. A strategy for BOOTH.

effects of longwall mining on shallow water KADNUCK. Cottonwood Mines.. Morgantown. (ed.S.S. J. Engineers.R. Report of Investi. MJ. MEHNERT. R. D.R. N. 1991. D. Proceedings of the Third Workshop on Surface MEHNERT.E. Utah. A Modular JOHNSON. & KAWAMURA.W.. A. 28. Series XII. B. F. Nottingham. BAUER. University of Nottingham. W. & TREVITS. Does longwall deep longwall mining. In: Mine Subsidence Control.F. M. 189–196. D. In: REDDISH. Influence of topography on the Three-dimensional Finite-difference Ground-water effects of longwall mining on shallow aquifers in Flow Model.F. West Virginia. 204–213. R. J.J. 786–793. In: PENG.A. MOEBS. F. Restoration.-M. 1187–1199. International South Africa. D. P.D. 1997. BRUTCHER. & PRESS.L. Morgantown. HOLLA. 2000.. 34. June 7992. Uni- Mining Science.L. D. D. J.S. J. L. West Virginia 3. Chapter A1. I.J.M. The effects of 1991. B. Geological Survey. & LIU. 1991. Hydrological disturbances Liu. G. Wasatch Plateau. Water- Morgantown.A. SENDLEIN. & GIBBENS. S. 26 HOLLA.M. (ed. In: NORTON. Subsidence Due to Underground Mining. US Geological Survey.B. (eds) Proceedings of the Third Workshop on HVORSLEV. 1985.M. MATETIC. Leslie County.M. 9405. Morgan- Pennsylvania. 34.V. I. June 1987. J.. MI. 197–203. &BUIZEN. International Journal of mining have a detrimental effect on shallow ground Rock Mechanics. West Virginia. US Bureau Longwall Coal Mining at the Deer Creek and of Mines Information Circular. WV. FRIEDMAN. S. HILL. Morgan. Information Circular. D.J. VAN ROOSENDAAL. W.A. S.A.. Ground Water Monitoring Review. Kentucky Geological Survey. 88.G. June 7992.S. Congress.. Effects Geophysical Methods. Ground Water Monitoring. Sedimentology & Recla- International Journal of Rock Mechanics and mation. May PAUVLIK.233 – 244. In: Proceedings of the National Symposium groundwater flow in response to longwall mining. R. DEMARIS. 1988. J. T. PARIZEK. 1992. F.) Proceedings of Basin. & PRICE.M. 1992. M. 147–158. Permeability of Coal measure Rocks Ph. L. WV. IAHS Publication 200. AGWSE (Association of of Longwall Mining on Hydrology. & SECRIST. Australia.. L. 1994. & BARTON.M. longwall mining subsidence on the groundwater con- KIM. & TREVITS. 1966. 1994. The Origin of Saline Formation Waters LEAVITT. 1991. National Outdoor Action Conference on Aquifer HUTCHESON. Hydro. Ground Water Scientists and Engineers) Ground Kentucky: Part 2: During Mining Conditions. West Virginia. BU 145. Response of Springs to sources. 138-143. CAREY. B. Reprint 30. KELLEHER. Effects of III: Calcium Chloride Waters. MCDONALD. & OWILI-EGER. Water Management Book 5. 1987. Transactions of the Geological Society of desaturation around mined panels. Salt Lake City. J. and Geomecha. 1980. D. December 1987. K.J. R. . West Virginia Report of Subsidence Investigations at the Rend University. 207–217.A. K. on Mining. N. Jefferson County. 1985.I. Techniques of the Appalachian coal field. & HARBAUGH. Dissertation. HODGSON. Appalachian Plateau.. 1139-1152. Utah. IMSRP-X (plus Appendices geological environment of full extraction mining. NIKRAZ. 33.J. published as ISGS Open File Report. & ELSWORTH. longwall mining effects on water wells. Stress relief fracture of longwall mining on ground water supplies. A. Short-term Pittsburgh. In: PENG.. and L. IMWA. (Society of Mining.J. L. Effects of Mining Subsidence on the Bureau of mines. WV. Illinois State JOHNSON. 90-141. US NEATE.M. M. Bulletin 36. Time Lag and Soil Permeability Surface Subsidence Due to Underground Mining. 1992. sium on Land Subsidence. Series X1. H. J. In: Proceedings of the Longwall USA Conference. September 1994. Mining Science. & VESELIC. water sources? In: Proceedings of the Fifth nics Abstracts. KlPP. MEENTS. Case study of Wales. Hydrology. The impact of deep (ed. & BAUER. (ed. West ways Experiment Station. US Army Corps of Virginia University. L. DlNGER. Topographic influence KIPP.B. Hydrologic variations due to longwall mining. D. BOOTH ELSWORTH. D. in Ground-water Observations. Illinois. SME (eds) Proceedings of the Fourth I. 1983. 397. R. Vicksburg. ting to mining under water bodies in New South MATETIC. & ESLING. ditions of a shallow unconfined aquitard in southern Evaluation of fully-coupled strata deformation and Illinois.A. P.R.) Land Subsidence: the Fifth International Mine Water Congress. N. 1991. & DINGER. September 1991. EVANS.1. Circular. Three-dimensional associated with increased underground extraction effects of hydraulic conductivity enhancement and of coal. 541–544.R. 1997-7).C. Overburden deformation and hydrologic changes Effects of mine dewatering on poorly consolidated due to longwall coal mine subsidence in the Illinois sandstones. R. Illinois. Annual Meeting.P. 143–153. M. 1C 9042. GRAF. town. Proceedings of the Fourth International Sympo.) Water Resources Investigations. In: JOHNSON.44 C. UK. Springfield. A.D. Morgantown. C.. Journal of Rock mechanics and mining Science. 13–24. 1987. June 7992. Lake Site. Pennsylvania. 1994. C. & SHIMP. strata fracturing and changes in permeability due to MATETIC. Illinois State longwall coal mining on rural water supplies and Geological Survey.J.J. 1997. I. J. S. versity of Kentucky Bulletin. 1990.) Proceedings of the Third Workshop on Surface mining on an overlying aquifer in western Subsidence Due to Underground Mining.A.. 228–236. 1. 1995. Kentucky Geologic Survey. 195-204.S.. M. University.. 1991.L. In: PENG. & TREVITS. Metallurgy and Exploration) Lubljana -Portschach. gations 4.J.A. 1951. Book 6. J. VAN ROOSENDAAL. February–1 March. stress relief fracture flow systems. A. The ground movement. M. Texas. Preprint No. & ELSWORTH. Some aspects of strata movement rela. Final town. Houston.T. J.S.A. S. 1990. control of ground water movement in the Ground Water.

KENDORSKI. 72-89. Environment. In: AIME Annual Meeting. B. YAU. August 1987.J. S. R.. . Illinois. subsurface mining. ALD. 2nd ed. In: Proceedings of Effect of longwall mining on ground permeability the First Conference on Ground Control in Mining. Water Resources cations from the Illinois Mine Subsidence Research Investigations Report. WERNER.N. Case Study of the Effects of well yields in the vicinity of room and pillar mines. 126-137. B. Virginia University. (ed. BRUTCHER.. P. Illinois and MURA. OFS 1997–9). of geophysical logs to estimate water quality of B.. D. Saline County. 3. R.. POOLE.B. town. impacts from underground mine subsidence in the N. J.. J. WAHLER & ASSOCIATES. H. 1987.S. Dewatering Lousiana. WV. & FOR. 9137. Overburden deformation and hydrologic changes Ground Water. Metallurgy and Exploration.. Eastern Survey. 1995.. Southern Illinois University.S. W. K..T. Littleton. A. R. Eastern TRENT. Hydrology of Area 25. Region.. edn. IMSRP-XI. 237–243.E. N. L.S. CAIN. mental Research Laboratory.) Proceedings of the Fourteenth Nottingham. & NEATE. G. MEHNERT.. 1-3 August 1995. 1981b. Final Report of north central Appalachians.F. G. West Virginia. Coal Mine Ground Control.M. RYAN. 88-4006. Proceedings of the Bureau of Mines Exeter Conference. RYAN. Subsidence Investigations at the Galatia Site.) Proceedings of the Third and underground coal mining.. Survey. PEART. Society Occasional Paper 9. CA. Francisco. Morgan- Report. Morgantown. West Virginia University. Miller Freeman Publications. G. In: RAUCH. D.C.L. US Environmental SCHMIDT. KELLEHER. February 1979. (ed. US Geological Survey. K. Chapter 2. Region. D. IL. & RAUCH. 1986.01-2. Morgantown. John Wiley. Industrial Environ- 79-132. & BAUER. BHS (British Hydrological Society) chanics. S. K. Eastern Mineral Law Foun.E. In: West Virginia Conference on Ground Control Problems in the Ground Water 1987: Status and Future Directions Illinois Coal Basin.W.. H. 1994. 81-636.S. (ed. X. CARTWRIGHT. US Bureau of mines ZUEHLS. D. S.O. Interior Coal Province. Ground Water Geological Survey Water Resources Investigations Manual: A Water Resources Technical Publication'. New Orleans. terization. KAWAMURA.A. DEMARIS.D. US US Department of the Interior. Water Resources Investigation Subsidence Due to Underground Mining. Morgan- SCHMECHEL. (plus 12 Appendices pub- 2. lished as ISGS Open File Report. 1994. SINGH. In: PENG. 1987. 1996. D. BENBIA. In: Hydrology in a Changing Appalachian coalfield.B. West Virginia.F. 80-82. Pennsylvania. & SANTY.A. Use VAN ROOSENDAAL.J. Pittsburgh. W. H.) Proceedings of the Third Work- Ground Water Sources in the Northern Appala- shop on Surface Subsidence Due to Underground chian Coalfield.. B. G.). F. due to longwall mine subsidence in Illinois. ElCHELD.J. US Bureau of Mines. IMWA. Eastern Coal Mine Geome. Virginia University.A. IMSRP-12. Study of YOUNGER.. D. (eds) Mine Drainage: Proceedings STONER. BAUER. Modelling of caving over the J. 1979. dence Special Institute. John Wiley. W. Illinois State Geological dation.. (rev. S. Denver. D. D. Y. In: Coal Mine Subsi. New York.31. Illinois Mine Subsidence ZUEHLS. Probable hydrologic effects of of the First Mine Drainage Symposium. 1998. Ground-water Hydrology of mine subsidence on shallow ridge-top aquifers in Marshall County. the Fifth International Mine Water Congress.) on the Effects of Longwall Coal Mining'. 1989. In: PENG. 81-403. D. 27. W. Open File Report. 1992. Morgantown.S. Pennsyl. 105–124. PL. Kentucky. Hydrological consequences dewatering effects at an underground longwall of the abandonment of regional mine dewatering mine site in the Pittsburgh seam of the northern schemes in the UK. June 1992.S. town. 1981. British Hydrological Technology Transfer Seminar. Preprint Society for Active Underground Coal Mines: Technical Mining. R. P. Morgantown.B. VAN ROOSENDAAL.J. RESTER. Findings and Practical Appli. Ground water impacts from surface CHUGH. West Automated data acquisition for subsidence charac. PEART. MEHNERT.P.J. 1983. CO Aspects and Cost Effectiveness'. West Virginia.K.L. West Virginia SINGH. Longwall Mining Induced Subsidence on Shallow In: PENG. 1990. West Virginia.a case study.. A summary of the ground water VAN ROOSENDAAL. XXV. OH. with Emphasis northern West Virginia.D. & LEAP. southwest Illinois. 73–82. 1. water and waste impoundments.L. & FITZGER- Information Circular. Hydrology of Area 35. W.K. In: REDDISH. Report of Mining. Strata University. J. Effects of coal SCHULZ. Carbondale.J. 1988. 76-88. and subsurface drainage. US Proceedings of the Third Workshop on Surface Geological Survey. 161–183. West Virginia.) Proceedings of effects of longwall subsidence . 1992. Factors affecting residential water Protection Agency. EFFECTS OF LONGWALL COAL MINING ON AQUIFERS 45 PENG.J. 1979. F. September 1994.E. No. San Review.A.W. 1998. Cincinnati. & KENDORSKI. Application of mechanical and ground- Lingan and Phalen mines in the Sydney Coalfield.P. (ed. V. M.B. WV..O.J. water-flow models to predict the hodyrogeological Cape Breton. (ed. Interior Coal Province. 1981. basal Pennsylvanian sandstones. 1985-1993. 9198. New York. vania. In: ARG ALL. Ground Water Monitoring Colorado. Pittsburgh. Program.A.. 1979. Water and Power Resources Service. 682–689. D.N.S. I98la. E. disturbance prediction for mining beneath surface WHITTAKER. & DEMARIS. BRAWNER. June 1992. REDDISH. F. 1998. WALKER. Carbondale. C. & PADGETT. Illinois State Geological ALD. Conference on Ground Control in Mining. R. Illinois. 1989. E. July 1981. 19 November 1986. RAUCH. & HEMPEL.T. P. West Investigations. J. 244–252. & KAWA.J. E. IL. & FITZGER- Research Program. TIEMAN. 252–260. C.

.

as has been shown by the examples of Picher. Fried 1972.g. 198. & ROBINS. 2000). 1993. 1987. active. es.g. be used in addition to quantitative analyses of • remediation strategies (Aldous & Smart rock stability in flooded mines (Hunt & Reddish 1987. In many cases. Potential tracers are restricted due to the complicated chemical composition or low pH mine waters. (eds) 2002.g. Lewis et al. Europe) polluted mine water objectives: must be treated. Gustav-Zeuner-Strasse 72. UK (Hamilton 1982. Germany (e-mail: c. D-09599 Freiberg/Sachsen. 1960. Mine water tracing CHRISTIAN WOLKERSDORFER TU Bergakademie Freiberg. Scott & Hays 1975. connections from the surface to the mine. 1996.de) Abstract: This paper describes how tracer tests can be used in flooded underground mines to evaluate the hydrodynamic conditions or reliability of dams. In some remote areas the company numerical models for performance pollution of surface or underground waters has assessment at potential radioactive waste dis- been taken to be of no serious account. remediation method. treatment techniques at the correct time after Abelin & Birgersson 1985. Special Publications. Germany (Gatzweiler Lachmar 1994). and the results of 29 mine water tracer tests indicate mean flow velocities of between 0. Cacas Techniques must be evaluated prior to ap. Hundreds of mines have been closed.7mmin. Wolkersdorfer land 1999). 1992. 1997) and to introduce passive • underground waste disposal (radioactive. Finally.00 © The Geological Society of London 2002. Aljoe & Hawkins 1997). Parkhurst ronek & Zmij 1977. A new injection and sampling method ('LydiA'-technique) overcomes some of the problems in mine water tracing.L. 1986. Brewitz et al. Reznik 1990. Lewis 1990. Cornwall. A successful tracer test from the Harz Mountains in Germany with Lycopodium clavatum. Sweden. ner & Williams 1993. Lee 1984). Lehrstuhl fiir Hydrogeologie. USA (Sheibach et al 1982. 0305-8719/027$ 15.1994. Himmelsbach & Wend- Fernandez-Rubio et al. and experience with mine water tracing is limited. microspheres and sodium chloride is described. 47 Geological Society. or vice versa. Davis 1994a. 47-60.S. its reliability published on the Stripa mine and the tracer must be proven (Younger & Adams 1999). after the since 1980–numerous papers have been chosen method has been applied. the groundwater table was ground surface and the mine.l . and to support remediation plans for abandoned and flooded underground mines. the mine owners have to find new. • optimize mining strategy (Adelman et al. plying natural attenuation or passive in situ 1992 (studies in the STRIPA mine.3 and 1. Wheal Jane. Mine water treatment. . et al 1997) or Konigstein. tracer tests in the vicinity of mines economic or ecological reasons. N. Mine water tracer tests are conducted in order to evaluate the flow paths of seepage water. remediation methods. & Meyer 2000). but in posal sites (e. Galloway & Erickson 1985. b. tracer tests can also • subsidence (Mather et al.wolke@tu-freiberg. Birgersson et al. London. Sawada et al. • mine water inrushes/inundations (Skow- Oklahoma. conditions within a flooded mine to minimize • underground waste disposal (non-radio- the pollution load (e. Furthermore. and tests conducted in the fractured crystalline tracer techniques are a useful tool to test the rocks). mine flooding (Younger 2000). United mines can be grouped on the basis of their States of America. Goldbrunner et al. Kirshner 1991. P. pecially if acid mine water is involved. Kirsh- fore. et al. Wolkersdorfer From: YOUNGER. Mine Water Hydrogeology and Geochemistry. There are only a few descriptions of successful tracer tests in the literature. 1988). There. in the recent are conducted to find connections between the past. can be extremely expensive. Wittrup et al. 1990. 1969). allowed to rise again as the underground Typical examples are the tracer tests that ac- workings flooded. 1985. Doornbos 1989. Wu et al. Miller & Schmuck innovative tools to control the hydrodynamic 1995). due to Usually. All tracer tests in most of the developed countries (e.

Iron Mountain: pH –3. Home et al 1987. stated by Davis (\994a. a flooded mine can be looked contaminating the mine water above the injection on as a karst aquifer in a conceptual model. be found in Wolkersdorfer (1996) or KaB (1998).48 C. Wirsing 1995). less experience exists and the are usually high (Banks et al 1997) and conse- expected results of an individual mine water quently limit the number of tracers that could tracer tests cannot always been found.g. Because Skowronek & Zmij 1977). effectiveness. In the case of fluorescein.g. and point. holes (e.g. The basic be used successfully. investigate the hydraulic parameters or inter. Most of the techniques are well described and. sites. free chloride (dams). adits. little can be said about tracers that are unsuitable connections of groundwater flow (KaB 1998). Lachmar 1994) or they used bore. As. raises and stopes. Canty & • estimating the decrease or increase of Everett 1998). wooden supports. at the surface to flow towards the mine through Liedl & Sauter 2000) might also describe the fractures (e. hydrodynamic situation in underground mines. as already (TDS). considered further. ground or surface waters and the mine (e. Wolkersdorfer et al 1997.g. Tracer tests have also been conducted to study Historically. The chemical composition of mine waters very Published results of tracer tests in abandoned often tends to be extreme: total dissolved solids underground mines are not common. A detailed discussion about the • investigating hydrodynamic conditions.g. KaB 1998) and investigate the more complex hydrodynamic energy mining is usually carried out by the use conditions was conducted in 1995 (Wolkersdorfer of boreholes only. Therefore. Habic & Gospodaric 1976. Wolkersdorfer & Hasche 2001). methods can be chosen. karst aquifers (e. Even conservative tracers. Diaz 1990. Laboratory tests using the chosen tracer and • investigating mass flow. Therefore. the tracer techniques developed for et al 1990). Anderson 1987. Gulati mines were simply to reveal connections between el al 1978. therefore. contaminants. usually only studies where they are commonly used to successful tracer tests–if at all–are reported. Galloway & Erickson 1985.g. strengths and weaknesses of tracers • tracing connections between mine and would go beyond the scope of this paper but can surface.g. Aquilinaetal. which easily adsorbs to organic materials. Up to now. high • testing the effectiveness of the bulkheads suspension load. Cacas Furthermore. the first tracer tests conducted in 'heat mining' in geothermal projects (e. WOLKERSDORFER 1996.g. In future studies. 1998). Most tests. This was Underground mines consist of a number of mainly due to the fact that no suitable method was shafts. in mine Nordstrom et al 2000) or metal concentrations water tracing. Kimball et al (1999) have already provided Doornbos 1989. KaB 1998) should be appropriate also for flooded mines.g. tracer tests should become a Many tracer tests have not been published in prerequisite for the evaluation of remediation the literature or are only available as academic strategies used for reclaming abandoned mine theses (e. • clarifying water inundations. these studies will not be 1996). radicals). aims are as follows: such as fluorescein. pH (e. only a small evaluate the environmental impacts of mine number of tracer tests have been conducted in effluents to watersheds. low pH. Wolkersdorfer et al Possible mine water tracers 1997. a range of tracers or ones.6. for mine water tracing. b). flooded underground mines to trace the hydro- dynamic conditions within the flooded mine itself (Aljoe & Hawkins 1994. which are similar available for injection of the tracer into the mine and comparable to the features found in karstic water at predetermined depths or without terrains. Kwakwa 1989. Randall et al 1990. injected the tracer numerical models of flow in karst aquifers (e. Bremerton 1989. This an example of how to use tracer tests and synoptic is often because they were either unsuccessful or sampling of trace metals in surface streams to the results confidential. there depending on the aims of the tracer test and the are both successful tests as well as unsuccessful hydrological situation. One of the first tracer their use is similar to studies in fractured rocks tests in a deep flooded underground mine to (see Himmelsbach et al 1992. Aims of mine water tracer tests Several classes of tracers have been used in Tracer tests are well established in groundwater mine water tracing (Table 1). typical mine water compositions are essential for . might be unsuitable under certain mine water conditions (e.

a each connected by two underground adits but detailed hydrogeological investigation is needed with different water chemistries (Table 2). Wolkersdorfer 1996 Carbondioxide Kirschner 1989 Tritium Parsons & Hunter 1972. Aldous & Smart 1987. Canty & Everett 1999. Mine water commonly flows with the procedures necessary for a mine water tracer mean velocities of 0. Davis 1994a. 1969. Located in the eastern Harz mountains. All tracers have to be and inappropriate selection of injection and selected on their expected or known behaviour in sampling sites. mine water. the dorfer&Hasche2001). based on the test of the expected residence time. Wittrup et al. Artificial and natural tracers that have already been used in mine water tracing Artificial tracers Salts Chloride (Mather et al. Aljoe & Hawkins 1993. Kirshner & Williams 1993. 45 days of laboratory tests. and preceding the StraBberg Case study: the StraBberg–Harz tracer test. Wolkers. Klotz & Oliv 1982. to clarify the most suitable injection and sampling The surrounding area of the mine consists points ('conceptual model of test site'). Pujol & Sanchez Cabeza 2000 Stable isotopes Adelman et al. the duration of the test. 4 weeks of laboratory tests with regular underground mine tracer sampling and visual analyses were conducted to test the stability of the tracers within Description of the mine the mine water (Wolkersdorfer 1996. Wu et al. Preceding the Nieder- schlema-Alberoda tracer test. To understand the aims of the test. Reznik 1990. 1986.3–1. Diaz 1990. StraBberg (Biwender). which can be used for a rough calculation Some conclusions are drawn. 1969. Parsons & Hunter 1972. that chosen. have been unsuccessful due to a lack of are partly influenced by metamorphism of the knowledge about the hydrogeological situation Ramberg pluton (intrusion age: 290 10 Ma). the cost of the tracer mannsberg. b) Rhodamine WT (Canty & Everett 1999) Solid tracers Lycopodium (Wolkersdorfer 1996. Wolkersdorfer 1996 Conductivity Bremerton 1989. 1992) Iodide (Wu et al. a short dence interval of 29 tracer tests investigated. Uranine (Mather et al. 1982. 1994) Sulphur hexafluoride (Kirshner 1991. Kirshner 1991. An important consideration when In the following sections a tracer test in a conducting such tests is the duration of the flooded underground mine will be used to explain laboratory test. see situation at the time of the tracer test is given. 1992) Borate (Lewis 1990) Dyes Fluorescein. Goldbrunner et al. test. and other tracer tests conducted or tracer in the mine water must be at least as long as described in the literature. Aljoe & Hawkins 1994. description of the mine and the hydrogeological excluding the maximum and minimum value. even if a suitable tracer has been the northern part Lower Carbonian rocks. 1995 Zinc variations Bremerton 1989 Temperature Anderson 1987. In tracer tests. In addition. The stability of the described. 1960. Table 6). Many predominantly of Lower Devonian rocks. 1997) Microspheres (Wolkersdorfer & Hasche 2001) Radioactive tracers Lorenz 1973 Neutron activation analysis Jester & Raupach 1987 Natural tracers Lead isotopes Horn et al. Diaz 1990 obtaining positive results. 1982) Rhodamine B (Skowronek & Zmij 1977.7m min – 1 (95% confi. Glasebach) material. Kirshner & Williams 1993) Lithium (Wu et al. MINE WATER TRACING 49 Table 1. and the cost of analysis. . Wolkersdorfer & Hasche 2001) Bromide (Doornbos 1989. Goldbrunner et al. StraBberg fluorspar mine is divided into three The most suitable tracer is chosen on the results mining districts (from north to south: Brach- of the laboratory tests. Wolkersdorfer et al. 1992.

StraBberg fluorspar mine. lakes and the groundwater. . arsenopyrite. The natural hydrogeological 1998). sphalerite. arsenopyrite. wolframite. polymetallic mineralization of the vein structures Before its closure. Abandoned Mines Ltd. karstic features and galleries (Wolkersdor.1 Li: <0. Galena. Map of Germany with the location of the StraBberg mine in the eastern Harz mountains and its main galleries and shafts. Pollutants are carried off 30km south of Quedlinburg and 6km west of the mine site by rainwater or drainage water and Harzgerode. scheelite. 1).g. important producer of fluorite in the former GDR Mine flooding is believed to be the most economic (Mohr 1978). the StraBberg fluorspar mine comprises several ore minerals of Permian–- was the largest fluorspar mine in the German Cretaceous age (e. Kuschka & Franzke 1974). The country (Gesellschaft zur Verwahrung und Verwertung rock has a permeability of about kf = 10. NO3: <0. Besides fluorite. Company for Remediation and Utilization of faults. pyrite.50 C.1mgL . Situated in the Mid Harz Fault Zone of the situation has been substantially impacted by the eastern Harz Mountains (Fig. Kuyumcu & Hartwig fer & Hasche 2001). Mean composition of the mine water in the Strafiberg mine during the time of the tracer test (30 Ma\ — 27 July 2000) in mgL .6 m s -i von stillgelegten Bergwerksbetrieben mbH. Fig. WOLKERSDORFER Table 2.1 . and groundwater circulates through fissures. the hydrothermal method for redevelopment of underground mines. galena. owned by the GVV fluorspar and barite occur in veins.5mgL -1 Shaft n Na K Ca Mg Fe Mn Cl SO4 HCO3 F No 539 15 22 2 56 21 21 12 28 198 64 5 Fluor 11 15 2 140 29 22 6 17 387 77 8 Glasebach 9 14 5 178 32 10 13 17 385 184 7 crop out. economic and chalcopyrite. pyrite. environmental reasons caused the closure of the siderite. Democratic Republic. sphalerite. In 1991. approximately many decades of mining. the StraBberg mine was the most deposited into rivers. 1.

2. and the following chemical parameters of the water within the injection points were actually used (from north three parts of the mine respectively. Consequently. manganese and sulphate.8.8. They also predicted the long- mainly found in the deeper parts of the mine term quality of the drainage water and. 539 shaft. as well as the bodies (Table 3). six injection and four sampling points depth. respectively) clearly showed that • What are the speeds of the mine water? stratification within the water body was taking • Why did the stratification break down after place (Kindermann 1998. Whilst the 3.1997 24. MINE WATER TRACING 51 Mining was started more than 1000 years ago an intensive investigation the DMT (German when silver. Ultimately. increasing with points. Fluor shaft. Evidence for the stratification were differences in temperature. the water within the mine was flowing from the StraBberg pit. and Glasebach pit. From the eighteenth century until new adits (the 'three-adit system') to drain the 1990 mining focused on fluorspar. accompanying in situ berg and the StraBberg pit still effective? temperature and conductivity measurements • Does the sewage water from the Sipten- within the No. south to the north.. conductivity and After careful investigation of the hydrogeolo- metal-concentration between each of the water gical and geochemical situation. the Department of Hydrogeology (Glasebach adit) connects the StraBberg and at the Technical University Mining Academy Glasebach pits on the seventh level. to separate questions: the Brachmannsberg pit prior to flooding from the StraBberg pit. a possible site of the final treatment day shaft at the StraBberg pit in 1910 marked the plant was chosen (Riiterkamp & MeBer 2000). stratification occurred as has been the three most important deposits of the observed elsewhere (Wolkersdorfer 1996). 539 shaft and the Fluor shaft (310 felde brook flow into the mine? and 147 m deep. After No. with concentrations. a dam tracer test was designed to answer the following was constructed in the northern adit. 204 52 19 600 50 18 525 * mHN. 340 40 8 466 31 6 359 367 22 1 143 23 1 204 284 42 9 478 27 11 417 272 79 2 389 22 1 196 134. 530 shaft.5 km long Glasebachquerschlag At this stage. significant differences. Sinking the Fluor these results. The StraBberg underground pit decreased in the 1980s. • How does the water flow between the three On 31 May 1991 flooding of the StraBberg and pits? Glasebach underground pits was started. were chosen. and between Almost immediately after the three adits were 1950 and 1970 the VEB Harzer Spatgrube joined finished. Selected constituents of the mine waters in the Flour and No 539 shaft in mg l -1 before and after the 3-adit system taken in use (after Riiterkamp & MeBer 2000) Depth* Fluor shaft Depth* No.2000 Fe Mn S04 Fe Mn S04 Fe Mn S04 Fe Mn S04 c. Hav- StraBberg mining district by driving two deep ing analysed the new conditions. .2000 26. the 1. Freiberg became involved in the project and when the ore reserves in the Brachmannsberg proposed a tracer test in the mine. particularly accessibility of potential injection and sampling of iron. level. based on (Bartels & Lorenz 1993). Between • Is the bulkhead between the Brachmanns- July 1992 and August 1998.5 km long that the mine water quality could be substantially Nordquerschlag (northern adit) connects the improved by constructing a drainage pipeline Brachmannsberg and StraBberg pit on the fifth from the StraBberg pit to the Glasebach pit. meters above sea level (Kronstadt elevation). Siptenfelde seepage. the physico-chemical and proved to be unsuitable. which was good quality water.. No. they thought Glasebach shaft). Glasebach shaft Table 3. 539 shaft. especially in TDS. showed to south): No. copper and lead were the targets of Mining Technology) proposed to construct three the miners. Fluor shaft. Riiterkamp & MeBer the installation of the three-adit system? 2000). start of the last production period.1997 24. the consultant adits on the fifth and seventh levels (from north to and the mine owners came to the conclusion that south: Brachmannsberg pit.2. One of the injection points finally Furthermore. 539 shaft (mHN) frnHKH 25.

Owing to the volume of the mine. The high flow rates) and the conditions of the mine sodium chloride was detected by continuous water itself (e. . 1).10. depth in shafts. As needed to raise the conductivity above has been shown elsewhere.50 . different sampling techniques were used. 6.18* Glasebach shaft (4 m) Microspheres green. 530 shaft (ca. pcs. 20m) Microspheres orange. (Table 4). 539 shaft (92 m) Microspheres blue.5 g 5 June: 12. Fluor shaft (Login 1997) were selected to inject the tracers into the GmbH. the Uhlenbach brook (PIC GmbH.08 . 1 5 um 4 X 107 pcs 5 June: 8. Radioactive tracers would concentrations of tracers at the sampling points.9 g 5 June: 12. 539.g. pieces. fluorescent dyes seem background. Wolkersdorfer et al Gommern/Germany). Filter Table 4. 1 5 um 4 x 10 pcs 5 June: 12. contain billions of tracer particles that can be assuming a total mixing of the mine water readily detected. Niehren & Kinzelbach the filter system and. No. as with salts. Tracer sampling and analyses Because of the requirements of a tracer test in a Owing to the characteristics of the tracers.10. Injection points. continuous During the tracer test. used for analyses (1998) presented an on-line microsphere counter as described below. the and 4X 107 pieces of microspheres would be use of salts would have required large amounts required at each injection point to get reasonable (e. Munich/ cedures and filter systems of the LydiA-technique Germany). two flooded mine (rough underground conditions. thereafter. but only quasi-continu. the tracers of ously using filters that have to be changed 50–1001 of mine water were concentrated within regularly (KaB 1998). nearly 10m3 of water sampling of the tracer is necessary. injected as deep as possible within the No. (flow cytometer) for microspheres with a diameter of 1 um and a flow rate of up to 1 m l m i n .1 to be used in groundwater studies. 30 um diameter) calculated on the assumption that the recovery and microspheres (15 um diameter) were chosen rate of Lycopodium is about 2-7%. 539 shaft (Login GmbH. Owing to a pending patent no more details about Fluor and Glasebach shafts because the three pits the injection technique can be given here. high suspension load).52 C. It was important that tracers were without contamination and to sample the tracers. and even small quantities tracer particle per litre of water must be present. and injection times of the seven tracers used Injection points (depth) Tracers Quantity Injection time 3 Siptenfelde seepage NaCl brine 20m (6. a flow conductivity measurement at sampling points in cytometer was unfeasible. Unfortunately. 12 hour basis.11* Add 40 h to the times marked with an asterisk. (Lycopodium Apparatus'.g. were pumped through the three filter systems microspheres and club moss spores cannot be that were installed and the filters changed on a sampled continuously. Consequently. Gommern/Germany) and the Glasebach deep flooded mine at the predetermined depths shaft (EcoTech GmbH. have been a good choice.pcs 5 June: 14.18* Fluor shaft (247 m) Spores saffron coloured 279. to be unstable in mine water due to chemical and The calculations used to estimate the mass of physical reactions and. relatively tracer required are similar to those used in karst large amounts have to be injected. 5–10t of NaC1). 15 um 4 X 10. As more than one injection point was needed Tracer amount for the test.18* Fluor shaft (247 m) Spores malachite green 264. 1 5 um 4 X 10. Therefore.pcs 5 June: 9. the pro. To guarantee reliable results.2 t) 2 June: 9. as LydiA (Lycopodium Apparatus) opened approximately 40 h later.44* No. WOLKERSDORFER (Fig.13 Fluor shaft (247 m) Microspheres red. but their detection is In the case of the Siptenfelde Brook an amount of expensive and the German authorities do not 200 001 of saturated brine (c. They can also be spores (Lycopodium clavatum. only a multi-tracer test could be It was calculated that 500–600 g of Lycopodium conducted.2t of NaCl) was allow these tracers to be used in large amounts. Dyed club moss aquifer tracing (see KaB 1998). that of because in both cases multiple colours can be used microspheres is 50-90% and that at least one at the same time. are connected to each other by adits at the shaft's deepest points.15 No. Bonn/Germany).

stored in 500 ml brown glass bottles. Each peak exclude any kind of contamination during sample is seen to start quickly and tail out slowly (see preparation. gene plastic filters were used for membrane After some rainfall events (Fig. an aliquot part of the whole filter was for collecting the solid tracers (microspheres. after at least 1 day of reaction. the fluorescent shaft. An increase in conductivity filter system was changed and the filters were could only be detected at the Fluor shaft (Fig. 3 July. the Fluor KaB (1998). Considering the the laboratory. if used in chemically unbalanced shaft (30 May–31 July) discharged 3. 11 June. the filter unit and the cantly after about 1 day (e. Changes before 6 June are due to moving the conductivity probe upwards in the shaft by 5 m. Therefore. working tables in the laboratory were cleaned to 11 June. NY 15 HC. 539 shaft. were installed at the No. Edetic acid. In 39% (2. Gottingen). Fluor calculated. Hamburg/Germany) and installed 5– Results 10m below the water surface. hydraulic connection between the Siptenfelde Gottingen.1 microspheres and the spores were counted under can be calculated for the meteoric and mine a fluorescence microscope (Zeiss.Smmin. During and carbonates. a significant change in conductivity that could be oxalic acid was added to remove both Fe-oxides attributed to the sodium chloride tracer. 9.34 pm). 539 shaft and Glasebach shaft) showed noticeable amounts of Fe-oxides. No. water flowing between the brook and the shaft's Fig. each with 100 and 15 um filters (NY Depending on the number of solid particles on 100 HC. Based on a distance of 2250 m After drying and mounting the 147 cellulose between the Siptenfelde seepage and the Fluor nitrate filters on glass plates.g. MINE WATER TRACING 53 systems. After each ductivity in the Fluor shaft increased signifi- filtration the Nalgene filters. . Nal. this recovery filters were carefully rinsed and the solids filtered rate is unexpectedly high. the geological and tectonic conditions. a mean effective velocity of l. None of the other sampling points (Uhlenbach Most of the 147 filter samples contained brook. 2).2t injected). Germany) 47mm in diameter. counted and the totals for the whole sample spores). inset in Fig. Plot of precipitation (Siptenfelde station) and conductivity in the Fluor shaft. shaft and Glasebach shaft. Kiel/Germany) the filters. suggesting a good through 8 |um cellulose nitrate filters (Sartorius. The arrow marks the time of the injection of the sodium chloride tracer into the Siptenfelde brook. 7. Every 12 hours the Sodium chloride. 10.2 m3 min~ quantities creates crystals that complicate the of mine water resulting in a total tracer recovery of counting process (Wolkersdorfer et al 1997).28 pm. as recommended by the time of conductivity measurements. 2) the con- filtering using a hand vacuum pump. 2). Mini piston pumps were used for sampling (Pleuger Worthington GmbH.4t of the 6. 2. brook and the mine. Hydro-Bios.49 pm.

5 days after tracer is a clear peak with 25 000 spores and another one Fig.54 C. A total of O^mmnT 1 (Table 5). and reasons for these peaks days decreased to nearly 4000 suggesting a other than the NaCl tracer can be excluded. and at the Glasebach shaft 11 days.2 3180 Siptenfelde brook Fluor shaft NaCl. The arrow marks the time of tracer injection. 530 shaft Glasebach shaft Microspheres 0. 539 shaft could be detected anywhere From To Tracer Velocity veff Distance (m min ') (m) No.5 observed elsewhere. WOLKERSDORFER Table 5. Therefore. From the injection point to the surface of the Club moss spores (Lycopodium clavatum). Breakthrough curves of the club moss spores detected at the Fluor and Glasebach shafts. 3. and in a relatively short time of 1.1 238 Fluor shaft Glasebach shaft Microspheres 0. A second peak with 6500 spores occurred 6 days after tracer injection. the spores have to flow 238m. 10. 3). the results cannot water flowing out of the three shafts. may have been contaminated from storage after Based on the ratio of the water pumped and the an earlier tracer test. 539 shaft.2 238 Fluor shaft Fluor shaft Club moss spores 0.5 days. .1 1773 No.3 4798 Fluor shaft Fluor shaft Microspheres 0.3 3180 Fluor shaft Glasebach shaft Club moss spores 0.1- and the Glasebach shaft (Fig. Club water. low hydraulic dispersion. the recovery be interpreted easily and it is not clear whether the rate is as high as 6%.820 in It is possible that some of the filter nets used the Glasebach shaft could be found after 8 June. No spores were detected at the No.2-1. 1000-6000 spores are due to contamination or Within the Flour shaft. Noticeable amounts of spores at the fluor shaft arrived 2.brine 1. Mean effective velocity of mine water in the Strafiberg mine. 530 shaft Fluor shaft Microspheres 0. 323. None of these peaks were injection.5 2250 outflow (Table 5).220 spores in the Fluor shaft and 200. the club moss spores not.5 days after tracer injection peak reached 199. after tracer injection. The abscissa shows the dates of sampling. thus the moss spores were only detected at the Fluor shaft mean effective velocity calculates to 0. No tracer from No.200 some 2. Nevertheless.

In the Fluor shaft. 530 and and Fluor shafts could be detected. independent of the location. The other peaks of microspheres from the Fluor shaft are negligible. with 7400 spores 3 days after tracer injection. with a decline to background within 2 Only microspheres from the No. The abscissa shows dates of sampling. the mean effective Some 13 days after tracer injection. be observed 2. the mean effect- Once again.3 m min-1 is indicated. 530 As the distance between the No. It cannot be Glasebach shafts is 4798m. injection to the sampling point.5 days. One day after the tracer injection 220 concluded that similar hydrodynamic conditions microspheres from the deep part of the Fluor exist in the mine.5 and 4 days later. Taking into consideration the two peaks sufficiently abundant to draw useful conclusions. but 13 days StraBberg mine: brief conclusions after tracer injection 3219 microspheres from the The general flow direction throughout the tracer No. Fluor and Glasebach amount of microspheres could still be detected.2mmin . 5). 9748 velocity is 0. a mean effective excluded that the LydiAs (Lycopodium Appar. and the shortest flow distance. the peak tails Microspheres. the velocities are shaft and the No. atus) lowered into the No. Breakthrough curves of the microspheres detected at the Fluor shaft. No. (max. 530 shaft reached the sampling point at the test was from north to south.2mmin -1 (Table 5). 530. including those shaft. 530 shaft days. 539.l–0. As already observed in the Fluor shaft. 4. MINE WATER TRACING 55 Fig. tracer detections during the StraBberg tracer test. shaft could be detected in the outflow from the shaft.1 . 539 and Glasebach Table 5 summarizes the results of all successful shafts did not open properly. were not pathway. 530 shaft arrived at the sampling point at the Glasebach shaft. microspheres from the Fluor From the results obtained. 530 shaft could be detected consistently around dmmin -1 and it can be (Fig. From the microspheres injected out slowly and even 3 days later a significant in the No. 4). the tracer had to flow 3180 m at the shortest injected into the Glasebach shaft itself. the maximum is reached very ive velocities are 0.2-l. as tracers injected Fluor shaft. only the microspheres from the No. microspheres from the No.: 3219 microspheres). quickly. velocity of 0. shafts. Between the injection point in the Fluor could be detected at the Glasebach shaft (Fig. shaft and the detection point in the Glasebach All the other microspheres. A sharp peak was observed that tailed out within 1. which explains the similar chemical . Based on the All parts of the mine are hydraulically well shortest distances of 238 and 1773m from the connected. The arrow marks the time of tracer injection. A significant tracer signal could still were never found north of their injection point.

99% of all the mean mea- tracer test's results confirm that the bulkhead at sured velocities are between 0. The arrow marks the time of tracer injection. Under of the mine and its surroundings.: 9748 microspheres).5 m m i n .001 and l l . Interestingly. the tests). 5. including the the current flow regime. . although The breakthrough curves clearly show that the former mine workers may have to be consulted to hydraulic dispersion within the flow path through assist setting up the injection and sampling plan. the reasons for that have not been fully understood. chloride tracer was recovered within the six From the literature review. there is a good connection between the Sipten. therefore. Tracer velocities depend on Furthermore. with the 3-adit system geological setting because tracers that are working. The abscissa shows dates of sampling. WOLKERSDORFER Fig. composition of the mine water in the Fluor and A critical prerequisite for a successful mine Glasebach shafts. and that the tracer is mine water can help to understand the flow transported after rainfall events only. injected close to the surface have to flow through not be re-established. the unsaturated zone. Because regime and trace elements might be useful to more than one third (39%) of the injected sodium support the outcome of the tracer tests. the partly unsaturated fissured aquifer and the In many cases. Future work. the hydraulic head and the hydraulic properties firmed the assumption that there is a connection of the rock. has to focus on: published. reasonably accurately from mine plans. Outlook Tracer tests in mines would benefit from the Tracer tests in partly or fully flooded underground development of new innovative tracer tech- mines are rarely conducted and even less often niques. Finally. but ongoing investigations will provide more details within the next few years. the sodium chloride tracer con. the northern adit is hydraulically inactive. it must be assumed that range between 0.1 . with a nearly log-normal distribution (24 tracer felde seepage and the northern adit. l m min-1 (Table 6).2 and 1. geochemical investigations of the drifts and shafts is small. previously observed stratification can. it water tracer test is a conceptual model of the test was believed that the general flow direction of site and extensive hydrogeological investigation the mine water was from south to north. Breakthrough curves of the microspheres detected at the Glasebach shaft (max.56 C. Until now. The mine geometry should be known between the Siptenfelde seepage and the mine. the experiences gained in the scientific world are limited until tracer tests in • tracer injection into predefined depths. • measuring of the exact injection time. flooded mines become more widespread. Previous to the tracer test. Therefore. water velocities weeks of the tracer test.

YOUNGER. BIRGERSSON. ment.-Prof. Lawrence Radiation Lab Report. C. H. 2.W.5 Wolkersdorfer et al 1997 mines. I.220 0. & GERARD.und Fluoritberg- Mansfeld GmbH and the GVV for financial and baus im Ostharz.5 11.1 Aldous & Smart 1987 P. W.077 0.3 This study Investigations. P. J. Mean of all 29 tracer tests: 0. Environmental I would like to express my gratitude to A. BACIGALUPI. VAUTE.L. P. suitable for mine water. Coeur d'Alene Mining District. JACQUOT. I would also like to thank P. R. • new tracers.736 1. E.L. Balkema.01 Canty & Everett 1999 Waste Disposal. details concerning necessarily reflect those held by the BST Mansfeld. the bad and the ugly. 1. 1993. BREWITZ. 0. & LORENZ. Report of 4. UK. WIDEN. would like to thank the staff of the BST Mansfeld.. 1994... & HAWKINS. financially supported by the BST Mansfeld. H.2 This study using fluorescent dyes. A... KULL. L. 0.. or any authorities involved.5 Wolkersdorfer et al. P. ARNESEN.077 0.R. 157–174.077 0. JEANNOT. JOHNSTON. 0.3 Parsons & Hunter 1972 172-178. 0. CONF 0. Kahmann.W. 1985.044 > 0.9 Wolkersdorfer et al 1997 GENTIER. BAUMGARTNER.E.8 Wolkersdorfer et al. W.01 Wolkersdorfer 1996 ADELMAN. In: COME. H. Use of Temperature Logs to 0.5 This study ANDERSON. 1987. 0. Stockholm STRIPA- group whenever help was needed. excluding the maximum and minimum value) References ABELIN.3 Aldous & Smart 1987 3-21. University of Idaho. 1-8. L.12 Canty & Everett 1999 (Colorado)'. BARIA. B. JUNG. L. & BIRGERSSON. Final Report on the Pinot Experiment 0. at the StraBberg mine.172 0.. 59.1-0. In: Proceedings of the 23rd Workshop on • tracer sampling and counting. 0.780 2. R..7m min-1 (95% confidence interval.773 0.6 0.4 Wolkersdorfer et al.2 0. 0.2 0.1-0. 1993.159 5.15 0.1-0. Reported distances and mean velocities of his efforts to make me publish my results! This paper worldwide tracer tests in underground mines. E. W. 0. ROSE. Swedish Nuclear Fuel the TU Bergakademie Freiberg supported the work and Waste Management Co.2 This study water movement in abandoned coal mined aquifers 1. Application of 0. BARTELS. Moscow. IVERSEN.776 1.. F. of BST Glasebach-ein Denkmal des Erz. 1985.. design and (km) (m min-1) instrumentation.14 Canty & Everett 1999 Livermore UCRL-6274.W.250 1. 1987. L.. J. & TEBBE. Urbana. B.171 0. Der Anschnitt. In situ 1999-00021). Tracer Migration Experiments in spent a lot of time taking care of us. AUDIGANE.. 180–190. (eds) Joint CEC/NEA Workshop on 0. 0. 1997 Nevada.0 Wolkersdorfer et al.W. E.3– 1. Tracing ground- 0. The table expresses the views held by the author and does not is given for comparison only. D.-J. 1997 aquifer testing in surface and underground coal 0. Washington.283 0. 32. Heinrich.35 0.180 0. A. 1992. Yucca Mountain.13 0. B. as well as K. Ground Water. T.T. & MULLER.723 7. Pools and Their Implications for Mine Abandon- nische Umweltforschung' (speaker: Univ. & NER- especially the Kohn family. 1997 Determine Intra-\vell-bore flows. US Bureau of Mines. R.4 Mather et al. who ETNIEKS.C. 1997. J.. Die Grube H. Geothermal Reservoir Engineering.. 1997 Mines. 45.780 0. California University. 26. C.7 Wolkersdorfer et al 1997 AQUILINA. Idaho Dr B.B.. MINE WATER TRACING 57 Table 6.. Rotterdam. 2. 1969 9468. M. I 144-158. personal support of the StraBberg tracer test. T. & HAWKINS.17 Canty & Everett 1999 ALDOUS. TRAN-VIET. R. Nottingham. AAGREN.5 1. 0.004 Aljoe & Hawkins 1993 in Underground Laboratories for Radioactive 0.001 Aljoe & Hawkins 1993* Design and Instrumentation of in situ Experiments 0.. 1969 Mine Water Congress. PL.2-1. for the improvement of the LydiA-technique. My colleagues at the Stripa Mine 1980-1991. S.7 0. W. BANKS... 1969 1960. 1989. Neutralization 3.1 Mather et al. & SMART. Migration Distance Mean velocity Author experiments in the stripa mine. Mine-water chemistry: the good. EVK1-CT. the EU Project PIRAMID (contract No. Bochum. Stanford SGP- TR-158. S.2 This study of Acidic Discharges from Abandoned Under- 3. Younger for experiments for the determination of macro- .J.6 Wolkersdorfer et al. the geological setting and hydraulic parameters are given GVV. & BANKS. A tracer test at * Result probably wrong. BRACH. the Soultzt-Sous-Forets hot dry rock geothermal site. in the literature cited. & MOMYER. Near-surface Acid Mine Water Graduate School 'Geowissenschaftliche und Geotech.. Master's Thesis. Also. In: Proceedings of the fifth International 0.229 0.3 Aldous & Smart 1987 ground Coal Mines by Alkaline Injection.798 0. Colorado School of 0. P. 1998.M. Golden.01 Aljoe & Hawkins 1994 8405165. F. the DFG BRETHERTON.6 Mather et al. 1997 ALJOE. IMWA 1. Hasche Geology.23 Canty & Everett 1999 ALJOE.216 0. This study had been TR-92-25. 2. Merkel) at the TU Bergakademie Freiberg and Master's Thesis.

J. Graz. HOTZL.. Transactions of the American of applications of non-radioactive but neutron Nuclear Society.W. Master's Thesis. Harvest MARINOS. A. KASS. J. & RAUPACH. RAUERT. In: Proceedings of the Annual einer hochsalinaren Porenwasserlosung–Laborver- National Meeting. 1998. American Society for Surface such und Modellierung. 1975. 1... Bern. 2™Pb) Hydrogeology of the Newfoundland Zinc Mine. MALOSZEWSKI. J. Synoptic Sampling to Measure Metal Loading cher Tracer in einem hochalpinen Magnesitberg. B. Investigations Report.W. 1990.E.. C. 1978. D. The flow model. Rico. RANK. T. 15. K I N D E R M A N N . Davis. K. Inter- gliche Stilllegung und Verwahrung von Uranerz. C. In: COME... 1999. 1997.D. WALTON-DAY.. P. E. J. D. Water Resources Institution of Water and Environmental Manage- Research. 1992. 25-27 July 1978. From Acid Mine Drainage.. G. Konferenz. 33-56. analyze the effects of reinjection into fractured FERNANDEZ-RUBIO. FEUGA. Modeling fracture flow with Jane. Balkema. W. ment. LEIBUND- a Mine Flooding Problem in Rico. & for the remediation of acid mine drainage in Wheal PEAUDECERF. Denver 91–2. HORN. Chartered and validation..C. 281–289. Passive treatment systems B. & ERICKSON. Colorado. Q. 289-302. as tracers for mixing of waters in the Konigstein Daniel's Harbour. Rotterdam. Uranium Mining and Hydrogeology. abandoned mineworkings in the UK. low. Stan- for controlling acid water in underground mines by ford University. J. (ed. V. L. Butte.L. 62-72. & MULLER. IAEA 623–633. 6 (3). M. Rotterdam. T. 26. 1976. Tracing Technique in Geohydrology. & STOURUARAS. Survey Open File Report. 1987.. Die Anwendung natiirlicher und kiinstli.C. Balkema. different aquifers (Investigations 1987–1992)– DAVIS.. P. 1990.A. Ljubljana. Water Resources bau. metalltransport in Sandsteinen unter Bedingungen ground coal mine.U. PROCTOR. DURAND. W.J. Colorado with a Discussion on the Use of Tracer RAJNER. J.. BENCALA.. 2000. J. Underground A. HALLET. W. & MEYER. niques in Water Resources Development. H.. Hydro geochemistry and Physical and thorogenic lead isotopes (206Pb. 1994a. P. The stability Budapest.L. implications of groundwater recharge upon shal- FRIED. JOHNS. TRIMBORN.7...2000. (1). A. LAMB. Preventive techniques Geothermal Program Review. ADAMS. Beitrdge zur Geologie der Schweiz. Berlin. DE MARSILY. International Journal of Mine Water.G.. MEYER. G. TSIAMBAOS.I. Denver 91-1..H. Germany. D. Geothermal Parameter beim Wiedereinbau von Reststoffen in . KOUKIS. GOLDBRUNNER. Institute Karst Research.J. 50. Princeton 15. & REDDISH. DOORNBOS. uranium mine. WOLKERSDORFER permeability and RN-migration in faulted rock Resource Council Annual Meeting: Geothermal formations such as the oolithic iron ore in the Energy—A Novelty Becomes Resource. H.C. B. Forest (Germany). CA. Transport phenomena in Report. P. R. R. 1998. Colorado Geological geologie. 103-112.. CANTY. KlMBALL. Newfoundland. Identification and Remediation of HlMMELSBACH. 11–14. activatable groundwater tracers. Wismat Guibit KASS. 207Pb.. 1985. 2. GULATI. Uranogenic DIAZ. & WENDLAND. (eds) Mine a stochastic discrete fracture network: calibration Water Treatment Using Wetlands.. ings Wismut 2000 Bergbdusanierung Internationale Wien. Hilo. 1982. (eds) Balkema. & EVERETT. In: Proceedings of St. In: BISWAS. W. R. Engin- a study in methodology.G. TILLIE. 192-193. 1972. 1995. Kontrolle geochemischer Tritium tracer survey at the geysers.C. & ZOTL. HOLZL. the Water Quality of the Acid Mine Drainage in the HORNE. Cornwall. P. Freiberg. 4. J. & geothermal reservoirs. & STROBEL.C.G. Konrad Mine Design and instrumentation in Hawaii.A.W. John's. Schwer- understand the hydrology of an abandoned under. The use of tracers to Technology. In: Proceedings of the fifth ESTEBAN ARLEGUI. M.C. S. JJ. S.. 1999.. 1994b.A. 520-529..A. M.. flooding. Rotterdam.) eering Geology and the Environment... 39-52. W. C. Colorado Geological Survey Open-File & VEULLIET.N. STICHLER.C. LIPMAN. MOORE. M. GUT. 237–239. P. 1997. P. UK. Geothermal Resource underground laboratories for radioactive waste Council. In: YOUNGER... G. Tracer test 3.N. 1998. In: Modelling of Water Resources Systems. Hydrogeologic Controls Upon Germany. Montana College of Mineral Science & J. 159-228. Washington. RAMSPACHER. R. E. JOHNSTON. B.. & GOSPODARIC. 31-36. Umweltvertra. BARBREAU. The Use of Tracer Dyes for the fractured rock test site Lindau/Southern Black Identification of a Mine Flooding Problem. E. Chemmitz 1–16. GALLOWAY. H. 1987. & CACAS. Dyes. In: Isotope Tech- GATZWEILER.. P.R. R. DAVIS. G. LEDOUX. JESTER. Water Tracing-Investigations in Slovenia 1972- ground laborataries for radioactive waste disposal. D. Montana. In: Proceed. G. 28. HABIC. R. H. House.A. & STIGER. 43. 1987. MOSER. T. Dolores County.. (eds) Design and instrumentation in under. 1989.M. MOSER. PL.E. & STICHLER. RUNKEL. M. K. GeoCongress. Techniques in Water Resources Development. CA 37–52. Steirische Beitrdge zur Hydro- Dolores County. 407-422. ZOJER. Sand Coulee-Stochett Area. 1–20. Miscible pollutions of ground water–.S. Newfoundland. Denver WRI 99-4018A Hydrologic. national Symposium on the Use of Isotope bergwerken-Fallbeispiel WISMUT.. S. HAMILTON. Mining and Reclamation. Freiberg 1. London. M. 2425-2430. 479-489.K. Use of tracer Injections and W. D. for evaluating nonpumping intraborehole flow in IAEA. Using tracers to HlMMELSBACH. disposal. S. FERNANDEZ-LORCA. E. Survey fractured media. Master's Thesis..58 C. & NINDEL.. H. Grundwasser. W. Montreal 2. Schlema. K. HUNT.

Journal of Contaminant national Biohydrometallurgy Symposium. 217-221. NIEHREN. J. Application of fracture-flow & LANYON. 1973. 108-123. 108. Jackson Hydrology.R.A. 7. (ed) National Symposium on Conference and Exposition. Deutsche Unsaturated Flow at Yucca Mountain PhD Thesis. N.C. Neuherberg. Idaho. 1984. Bureau of Mines Report of Investi- an in-situ Leach Cell in Abandoned Mine and Mill gations. & Warrendale 77-87. R. In: Inter. 1991.A. 2. 2000. United States. Ground-surface Delineation of between mining subsidence and groundwater Fractures That Appear to be Connected to occurrence at Aberfan. Proceedings of the Consultants 59.H.. 1998. 134Cs und l44Ce in drei typischen Environmental Sciences and Technology. Spokane RI-9583. the Kamaishi Mine.H. Geologie und Minerallagerstatten des Tracers. gations: Washington. Use of a Tracer For Moscow. MILLER. Bureau of Mines Report of Investi. D.B. Occurring Gaseous Tracer. wasser. M.. Engineering ground copper mine. HERSHMAN. 136-154. Picher Mining District.W. 1990. Harzes. R. Loading. 1997. Untersuchungen zur Tracer for Unsaturated Flow at Yucca Mountain hydraulischen und hydrochemischen Situation in Borate Adsorption by Some Volcanic Tuffs. 49. Idaho Superfund Site Using MOHR. M. 1037–1099. M. & OLIV. Northeast Japan. O. Knoxville. Mining Engineering. Journal of Hydrology. NORDSTROM. 85Sr. WILLIAMS. Charakterisierung von abandoned underground mines-water pollution Karstgrundwasserleitern durch Simulation der prevention and control. L. J. Inc. W. & MEBER. T. Environmental Pol- Final Report of the Camborne School of Mines lution. E. M. Controlling acid mine drainage from the Formations.M. Aufgaben der into the movement of groundwater from Bryn Pit GVV-Gesellschaft zur Verwahrung und Verwer. Performance assessment of nuclear REZNIK. C. OF 86-0481. Use of KWAKWA. 56. P. University of Idaho. F. 1994. 1417-1436. Evaluation of Borate as a Potential RUTERKAMP. H. YAMAMOTO. acidic Mine waters from Iron Mountain. Moscow. R. SCOTT. I. A.J. R. GRAY. D. Freiberg. 1993.E. 2000. The Minerals. 7791. Oklahoma. 1969. nuklide 125I. K. C. 71. above Ebbw Vale. .K. & KINZELBACH. Pergamon Press. Hydrothermalite des Harzes. RANDALL. 18-22. Metal 395-401. Washington. Project. & SANCHEZ CABEZA. Zur Kenntnis der Survey. R.. LEWIS. LTM Mining: A New Beginning. 9. 68-72. PARKHURST. Dry Rock: Geothermal Energy. Volcanic Tuffs.M. Water Pollution tung von stillgelegten Bergwerksbetrieben mbH. A. & TAKAHARA. UCHIDA. MATHER. LACHMAR. 1982. D. T. ALPERS.L. In: BARIA. F. PARSONS. 31-44. 43. Evaluation of jointing in hydrogeology to acid-mine drainage at the Bunker the carnmenellis granite. 196-201. 35. & FRANZKE. Bergbau. Open File Report.B. FlieBen. Schweizerbart.L. (ed) Hot Hill Mine. University of Idaho..I. 1993. Stuttgart. J. 1. 75-96. Y. In: In: Proceedings of the Eleventh Annual WATTec GRAVES.H. Non-sorbing TILK. 33.S.D. Phase 2B Ebro River waters (Spain). Kellogg. 568-572. Reston. 257-269. KIRSHNER.A. Hole. Berlin. H. Wissenschaftliche Mit. Hydro geological Evaluation of Research. J. 1974.J. Metals and Materials Society. Robertson Scien- 155. Geology. International Journal Mine Water. Aquifergenese und des Warmetransports. and Chemical Reactions. 1972. Inactive and LIEDL.. & HUNTER. Investigation KUYUMCU. Golden. of Evaluating Fracturing of Underground Oil Shale 1982.. Borate Adsorption by Some Kensington. & DOE. D. Oxford. in situ Stope Leaching Solution Containment KIRSHNER. R.N. Numerical simulation of bulkheads tracer migration experiments in fractured rock at to reduce uncontrolled discharge from an under. & WILLIAMS. R. Lexington. F.W. 1975. D. University of LEWIS. & HARTWIG. P. tritium to predict soluble pollutants transport in (eds) Hot Dry rock Geothermal Energy.. R290. 2000. D. Zeitschrift fur geolo.A. Heidelberg. Moutan-technologic (DMT). LEE. 1990. F. GSF-Bericht. South Wales. 125-149. J. Waste at the Kellogg. M. 9-16. Unpublished report. & GENES.-J. & SCHMUCK. Artificial geological evaluation of an in-situ leach cell in colloid tracer test–development of a compact on- abandoned mine and mill waste at the Kellogg.R. J. Control.E. Grund. Monmouthshire. SHIMO.E. SAWADA. 5. & HAYS.G.M. 1998. 1990. Hydro. paths using method of streaming potentials. Tracer results. BLOWES. & JENKINS. K. tific Publications. H.W. (eds) US Environmental Protection Agency. line microsphere counter and application to soil Idaho Superfund site using tracers.E. H. 1988. D5-D9. gische Wissenschaften. R. 2000. Mine-water Discharge.L.D. MCGOWAN.E. 1989. M. R. & SAUTER.E. 49. Colorado School of Mines. Sanden Norddeutschlands bei wasserungesattigtem 254-258. PTACEK. 1989. MINE WATER TRACING 59 ein stillgelegtes Bergwerk.. C. LORENZ. WILLIS-RICHARDS. R. 2000. Camborne.. den drei Teilrevieren der gefluteten Flussspatgrube Evaluation of Borate as a Potential Tracer for StraBberg: 46. 249-259.. Journal of Underlying Mined-out Openings Using a Naturally Hydrology. S.. Verhalten der Radio. Gelsenkirchen. Radioactive Tracer Pulse Method SHEIBACH. Determination of groundwater waste isolation systems Shaping our energy future. 34. Negative pH and extremly KLOTZ. NICHOLLS. G. California.L. US Geological KUSCHKA. J. use of tracers to investigate the relationship KIRSCHNER. In: MICHAEL & BAKER JR. DC EPA-440/9-75-007. W. B. The teilungen-Sonderheft.. column experiments. PUJOL. National Symposium on Mining. In: PARKER.

97.und Wasserver. KYSER. Tracer pochodzenia wody wyplywajacej zza obmurza test in the abandoned fluorspar mine StraBberg/ szybu znakowaniem rodamina B. the United Kingdom. 84-97. Regina. 6. Bristol. YOUNGER. Poland.. E. [Determination Harz Mountains. of Brine in Saskatchewan Potash Mines. vertical groundwater recharge locations in a coal WITTRUP. Q. 159-165.B. 1996. 1999. Przegl WOLKERSDORFER. M.60 C. R. Hydrogeochemische passive treatment technologies for mine waters in Verhaltnisse im Flutungswasser eines Uranberg. WOLKERSDORFER SKOWRONEK. 1-216. & ADAMS. 2001. I. mine. Wu. 8. Li. shaft lining traced with Rhodamine B]. Saskatch. T. YOUNGER. 1992. (ed. Quelle 14 der Freiburger Energie. 1977. Wissenschaftliche of the origin of water flowing out from behind a Mitteilungen. In: KRANJC. B. Tracer Beurteilung des Grundwasservorkommens und Hydrology. & DANYLUK.) Balkema..L. Special Publi.. 57-67. uranium mine using Lycopodium clavatum. Mine Water and the werks-Die Lagerstatte Niederschlema/Alberoda. Corn. In: International Symposium on Water The Use of Stable Isotopes to Determine the Source Tracing (SUWT). 2000. Y. & HASCHE. . Predicting Mine Clausthal. 19 (2). Ch. Unpublished report Geolo.. 13-19. Ch.L. Rotterdam. Freiberg. Okreslenie WOLKERSDORFER. Clausthaler Geowissenschaftliche Dissertationen. Y. Environment. D. Gliwice. 1995. 33. Ch. A. 1997. The adoption and adaptation of cation. ewan Geological Society. 1986. A. P. TIAN. 7. Environment Agency. Development of a tracer test in a flooded WIRSING. T. P. Water Rebound.K. Germany. G. & FELDTNER.. N. & XIA. 5Q. JIN. application of artificial tracers in determining gisches Laudesaut Buden-Wiirttemberg. M. WOLKERSDORFER. Balkema. 377-385. An sorgungs-AG: 45. TREBUSAK. 16. & ZMIJ. Abgrenzung eines Wasserschutzgebietes fiir die Rotterdam. 365-367. Hydrogeologische Untersuchungen im Grubengebaude des Schauinslandbergwerks.

co. NG17 2NS. followed by a period of more brisk recovery until the next large void was reached. 61 Geological Society. The result of this policy was a general lack of experience of mine water recovery and continuing doubt about underground connections. Sutton-in-Ashfield. Mine Water Hydrogeology and Geochemistry. N. that is governed by the hydrogeology of both berland. London. Hutkwaite. The 1990s.g.g. recovery follows an exponential curve water in the mine and the source of aquifers and with the rate of recovery reducing with time. monitoring of mine water recovery in data suggest that water inflow rates are probably several abandoned coalfields has shown that. 61–73. Adams & Younger 2001) to large the natural in situ strata and the changes to that From: YOUNGER. Estimations of mine water recovery made by British Coal at the time of these closures were generally based on a water inflow related to the volume of water pumped from a mine and a residual void-space calculation. The control of mine water during the period when coal mining was a nationalized industry was generally based on a safety first principle. The monitoring and modelling of mine water recovery in UK coalfields KEITH R. The paper concludes that modelling of mine water recovery. related to the difference in head between the in general. The void-space was calculated using roadway dimensions for supported excavations and a figure of 10% of the original extrac- tions thickness for unsupported (total extraction) workings (National Coal Board 1972). can be used to give reasonably accurate predictions of recovery times and flows. Banks Figure 1 shows the recovery curves for several 2001). Using this principle it was assumed that mine water recovery would proceed as a series of steps.uk) Abstract: This paper draws together the information that has been obtained on mine water recovery since the large-scale closure of coal mines in the 1980s and 1990s. & ROBINS. UK (e-mail: -whitwork@imcgroup. Sherwood & Younger 1997. Special Publications. It also implies a general interconnection mining units in the UK. following cessation of pumping. the area that has been dewatered (cf. p. mine water recovery follows precise exponential curves that appear to be independent of the distribution of mining voids. based on mine water inflow and estimated void space. Robins et al 2002). PO Box 18.L. (eds) 2002. The units vary from a of the mine workings within a block of ground single small mine (e. mine water recovery follows an exponential curve similar to the recovery of an aquifer following a pumping test. Nottinghamshire. Monitoring of mine water recovery interconnected areas or whole coalfields (e. 0305-8719/027$ 15. Since the coal mine closures of the 1980s and Burke & Younger 2000. The reasons for this type of mine water recovery are discussed and examples are given of the use of the data for both the interpretation and modelling of mine water recovery. Sherwood 1997. Whittle in Northum. 198. WHITWORTH IMC Consulting Engineers. The data show that. with very little recovery when water was 'filling' a large void. The large-scale closure of mines in the 1980s and 1990s mean that in many cases whole coalfields were abandoned and that the pumping of mine water either completely stopped or was greatly reduced. methods for the prediction of mine water inflow and recovery modelling are proposed and the problems associated with mine water recovery modelling are discussed.00 © The Geological Society of London 2002. Several previously un- published examples of mine water recovery data from around the UK are included in the paper and there is a detailed assessment of mine water recovery in the East Fife Coalfield in Scotland. but that water level monitoring is essential for precise predictions.S. This meant that when doubt existed about underground connections between modern mines and old abandoned areas of workings. mine water pumping always continued in the old areas. at least at large scales. . In coal mining areas where no water-level recovery data are available. The monitoring of mine water recovery by IMC Consulting Engineers on behalf of the Coal Authority (the government agency set up to look after the non-privatized areas of coal mining) has shown that. These curves are very similar to the recovery curves observed following an aquifer pumping test.

workings be allowed to flood. mine water recovery in the around long wall faces. Only where these blocks (e. has been monitored by the active working blocks of a mine were kept 'dry' Coal Authority since mine water pumping was to avoid sterilization of reserves until all the coal abandoned in 1995 and provides a good example of the general principles of mine water recovery had been extracted.62 MINE WATER RECOVERY. This recorded at the surface and water inflow interconnection of workings in coal mines could monitored underground. Nuttall el al 2002). Frances and Michael had known. older areas of workings level where the water recovery would not result has been abandoned for a number of years. situated to the north of The results of these regulations was that the the Firth of Forth. Younger et al 1995. active workings. MONITORING AND MODELLING Fig. Sherwood & Younger were isolated from the rest of the mine would old 1997. occurrence of several major inflows of water into Figure 3 shows the general layout of the coalfield mine workings during the 1970s and 1980s and the current monitoring sites. water-filled mine workings (National Coal Board 1956). be anticipated from the way both natural and Figure 2 shows how the natural stresses in the mining-induced permeability gave rise to the ground are altered around a typical longwall rules that governed the way coal can be mined in panel (National Coal Board 1978). as yet. 1. but latterly these areas were allowed to the recovery period did not generally occur. Initi- in a risk of flooding to deeper. mine water had been pumped for safety Detailed monitoring of mine water levels during reasons. Since the resulted in a renewed interest in water movement cessation of pumping. strata caused by mining (Sherwood 1997). The resulting research workings has followed the typical exponential tried to link the mechanical effects of total trend with. which can an aquifer (including unconsolidated material significantly increase the vertical and horizontal likely to flow when wet) or within 45 m of other permeability. The regulations state that the total resulting from these stresses cause the opening of extraction of coal is not allowed within 60 m of fractures and bedding planes. Sherwood 1997. The water levels up until 1995 (Younger et al 1995). These rules were brought in following East Fife coalfield experience of serious water inflows into mines when working too close to other bodies of water. and the recovery 'curve' was assumed to closed in the 1970s but were pumped to control be related to the filling of the void spaces. only minor deviations from the extraction coal mining with the subsidence data trend.g. and then only to a Mining in the shallow. only recover and overflow to the deeper mines of the start and end of the recovery period would be Michael and Frances. Mine water recovery curves from a range of major UK coalfields. The East Fife Coalfield. These deviations occur on an annual cycle . The strains the UK. ally.

R. by an area showing directions of water movement determined where no mining has been carried out or by from hydraulic gradients and mine plan evidence. Similar gradients of up to 1 in 1000 such as mudstones. A (adapted after Sherwood 1997) showing major bodies of mine workings (shaded grey onshore. different types of mining and varying lithologies. adjacent to the mine developed (see Fig. General layout of the East Fife Coalfield used to confirm the separation of mining units.5 km through the mine workings to Lochhead. such as sandstone. The shallowest areas. (0.4 m are caused by compression of the Coal Measures strata due to weight of water at high tide reducing the storativity of the aquifer by forcing water out of joints. The range blocks. adjacent to the workings. during spring tides. . Blyth where. artificial barriers such as dams. water levels in these blocks take up the of hydraulic gradients are probably related to the same recovery trend. The higher hydraulic gradients will currently developed in the East Fife Coalfield reflect workings that have closed or been vary from about 1 in 200 (0. WHITWORTH 63 Fig. The varying amplitudes of the fluctuations probably reflect the depth and extent of the under sea workings in an area. with arrows adjacent workings by Coal Measures. stippled where mining unit may comprise a single colliery or a under sea) and positions of mine shafts and boreholes group of interconnected collieries isolated from used for water level monitoring purposes. as with the mine workings or fractured high-permeability rest of the coalfield. Figure 5 shows the tidal variations in water level monitoring at Frances.0005) to 1 in 600 backfilled and have only low-permeability strata. can be Fig. Monitoring and mining connections Monitoring of mine water recovery. as well as showing interconnection of workings.5 m have been recorded.001) are found in most monitored coalfields The monitoring of water levels in East Fife is in the UK carried out using pressure transducers and data loggers recording a water level every 15min. The amplitude of the fluctuations of up to 0. although. The largest fluctuations of mine water due to tidal compression have been seen at Bates Colliery. The low hydraulic gradients will reflect open have yet to start recovery. 3.0016). a hydraulic gradient has strata. and correlate with annual variations in rainfall. 3). The hydraulic gradients workings. 2. Muriespot and Dalginch. These oscillations take some 4 h to travel 2. fluctu- ations of up to 2. monitored at Randolph. (0. Strain patterns around a typical longwall panel showing the zone of increased water flow and bed separation (after National Coal Board 1978). This allows very minor fluctuation in water level to be examined in detail. K. The similar hydraulic gradients in different Figure 4 shows that as the mine water recovery coalfields suggests that the various coal mines reaches the level of each of the shallower have similar hydraulic conductivities.

j. MONITORING AND MODELLING Fig. levels on either side of a dam has been static head. Since then the water level Fig. 4. Monitoring of water levels on the major mining blocks. Figure 6 shows only a limited water head was allowed to build up an example from Durham where the mine water against the dam. 5. dams were sometimes rise side of the dams has shown that the British constructed in roadways to separate the mining Coal policy of continued pumping was entirely units and prevent water migration. Where only single roadways connected construction. but were rarely tested because pump. These dams justified. as in nearly all instances where water were generally designed to withstand full hydro. Mine water recovery in me East Fife Coaleia. . as monitored in me snarts ana oorenoies marKea on Mg. East Fife. monitored. sometimes many years after caused the dam to fail.64 MINE WATER RECOVERY. The resultant head of mining in an area. Dams. The effectiveness of the dams recovery in an area isolated by a dam was greater was therefore only proved with the total closure than the rest of the block. and a high hydrostatic head has ing in the abandoned area usually continued and built up on one side of the dam. Tidal fluctuations in mine water levels measured at the Frances and Lochhead Collieries.

(This had always been suspected due horizontal stresses. The most striking example of this is of 80m head difference across the roadway the blockage of a 14x10 ft (4. the should have become harmonized. the rest of the mining block. has fallen and harmonized with the recovery in –180m OD (180m below Ordnance datum). constructed between Barnsley Main Energy at Monk Bretton has proved that the and Monk Bretton Collieries at an elevation of Barnsley Seam remains unflooded at a level . combined with swelling to a flow of methane from the Monk Bretton mudstone lithologies. The original monitoring borehole at Monk project by British Coal (1986). K. Table 1 shows a sudden rapid recovery. Figure 7 shows lithology around the opening and the stresses that as the water level at Barnsley Main neared acting on the opening. These water levels give a minimum blockage. Water levels were expected to recover at differing rates at Barnsley Permeability of mining connections Main and Monk Bretton until the connection The permeability of a mining connection varies at –180 m OD was reached when the levels greatly dependent on the type of opening. then a return to the some typical Coal Measure permeability values exponential curve precisely followed by the mine obtained from drill stem testing (D. indicating an overflow to deeper with a large pressure difference across the workings.) in water recovery below the — 180 m OD connec- boreholes for the north east Leicestershire tion. 6. as connections have permeabilities up to several would be expected from an overflow at a major orders of magnitude greater than the permeabil. In general. these 'closures' only occur or oxygen-diffident air. result in closure of a mine mine. connection. examples have been found where major borehole shows a static level of approximately underground roadways have become blocked — 100m OD.T. stop when mine water in areas of total coal extraction and the edges of levels rise above the insets or connections from these areas usually retain a 'high' permeability.26 X 3. However. In general. WHITWORTH 65 Fig.05m) blockage between Barnsley Main and Monk roadway supported by steel arches at 1m Bretton.S. Mine water recovery records for selected shafts and a borehole (Warden Law) in the Durham Coalfield to the east of the River Wear. the shafts to the workings. Based on personal Bretton (Lundwood borehole) filled by British observation of changes in flow it is believed Coal in 1994 has been redrilled to confirm that that permeabilities in mine workings can revert mine water levels at Monk Bretton have not to near natural permeability where vertical or recovered. The new Lundwood However. This roadway was driven specifically for the purpose of water drainage. A recent borehole drilled by Alkane spacing. mining — 180 m OD the recovery decreased rapidly.) The emission mine gases. within 6 months there was ities found in UK Coal Measures. R. either methane opening. showing the effect of failure of a dam on water levels in Sherburn Hill Colliery.

i. and the mine volume derived from the mine plans and recorded . Younger & Adams 1999. Mine water recovery modelling The modelling of mine water recovery is important for assessing the risks to aquifers from recovery of contaminated mine water. Figure 8 shows an example from Easington Colliery in County Durham of a projection compared with the actual recovery curve. The residual void left in areas of total extraction was generally considered to be 10% of the ex- traction height (National Coal Board 1972). the volume of water needed to fill this space over a given period can be calculated. This method has been used in the Northumberland and Durham Coalfields. The mining void is calculated by digitizing the area of the mine workings and using recorded extraction thick- nesses and roadway sizes. MONITORING AND MODELLING below the connecting roadway level of — 180 OD. and for the prediction of the timing and the flow rate of potential surface discharges (e. Monitored mine water recovery curves can also be used to estimate the flow of water into a mine and the likely volume of any surface discharge after recovery. It is also very useful in the assessment of coal mine methane (CMM) reserves.66 MINE WATER RECOVERY. Therefore. which gives a more accurate recovery in terms of time. A key part of any projection is the source aquifer. Figure 9 shows the calculated inflow for the East Fife Coalfield based on the monitored recovery. This figure was based on filling of areas of abandoned workings by known flows. Banks 2001. 200m head of water.g. Adams & Younger 2001). Sherwood 1997. Using mine water recovery curves Modelling can simply take the form of forward projection of a monitored exponential mine water recovery curve (Younger & Adams 1999). Areas of total extrac- tion and areas of partial extraction are recorded separately to allow the amount of compaction to be varied for the calculation of the residual void. This information is needed for the project to establish a theoretical maximum recovery level. Knowing the rate of recovery and the estimate void space. and has so far proved to be very accurate. as well as in East Fife. and the maximum head of water in the aquifer prior to mining. the Monk Bretton side of the roadway must be unsaturated and the head acting on the roadway is the difference between the level of the roadway (–180m OD) and the current water level in Barnsley Main (+20 m OD).e.

which were purposely connected by a major roadway at — 180 m OD before abandonment. the best-fit figures suggested that at Fig. 8. extraction thickness. Using the recovery. WHITWORTH 67 Fig. water inflow has reduced exponentially. the modelling was further 30000m3 day"1 inflow and matching the recor- refined by using the volume of mine water ded recovery. County Durham. 7. K. The lack of harmonization of the two recovery curves above this level demonstrates that the roadway has so thoroughly collapsed as to prevent flow from Barnsley Main to Lundwood. . Projected mine water recovery and subsequently monitored levels at Easington Colliery. The graph shows that the pumped (30000m3 day ]) as the initial inflow. with a Using this inflow and a calculated residual mine fairly rapid initial reduction followed by a more workings void (based on 10% of the total gradual decrease. The later inflow rates can be extraction) resulted in a recovery that was much used to assess the potential surface outflow after quicker than actually occurred. In this case. Mine water recovery at Barnsley Main Colliery and the adjoining Lundwood workings. R.

the residual void was in inflow generally equals outflow. which may subsidence would only be 35% of the extraction or may not be interconnected. 10. MONITORING AND MODELLING Fig. The conceptual model for residual underground voids are greater than 10% of an individual mine or interconnected block is the original abstraction. 300-350m below OD. with 65% of the extraction void being all the units can then be put into three basic underground (National Coal Board 1975). in many water inflow is required.g. A-D.68 MINE WATER RECOVERY. The various the order of 20% of the initial extraction volume. water inflows were collected and pumped inflow and pumping data from a mine can be to surface separately providing accurate data on used to interpret the steady state situation where these flows. as shown in Fig. 9. Water inflows into thickness. During mine water recovery. At 100m depth a 200m based on a general model. The categories. In this case the recorded cases. This water may originate from between the mine workings and bed separation surface superficial deposits.Adams wall abstraction would also suggest that the & Younger 2002). such as the Sherwood workings) is believed to be the mechanism Sandstone or Magnesian Limestone. Measured and modelled mine water recovery and derived values of calculated water inflow over time for the East Fife Coalfield. shaft . which is 90% of the extraction of mining in the UK and assumes four basic thickness. wide longwall panel can result in maximum The model is based on the general development surface subsidence. Coal Measures (which form a substantial part of the void above the aquifers or major aquifers. based on the depth in source aquifer (e. as most of the aquifers recovery. inclined roadways to access the Coal Measures through a major aquifer is rare due to the increas- ed length of drivage in the aquifer and the related Mine water inflow data increase in costs to seal the roadway. an alternative method of calculating are at relatively shallow depth and. Water Where there is no monitoring of mine water qualities are usually good. The water quality of the would suggest the British Coal figure of 10% is at various inflows can also be assessed at this stage the low end of the range. but the East Fife recovery vary during recovery. problem then becomes 'how much of this under- ground void is hydraulically connected with the mine workings'. inflows are then fitted onto a conceptual model The residual void percentage will vary with for each mine to assess how these inflows will lithology and depth. Sudden hydraulic connection Shaft water. using either actual analysis data or water quality Monitoring of subsidence above areas of long. However. at 500 m depth the maximum depth controlled mining units. Shafts controlling the large inflows of water noted both generally form the only major interconnection in the South Staffordshire and the Selby Coalfields between a mine and a major aquifer The use of (Whitworthl982).

Banks 2001. dependent on source. the pathway and the length of the pathway travelled by water. The Modelling recovery using mine water rate of water inflow to a mine from these sources inflow data will decline once the water recovery in the mine has reached the level of inflow. water. the minimum recovery period for a mine. water inflows from major aquifers will only start in Yorkshire with annual fluctuations in water to decrease when water levels in the shaft reach level related to variations in rainfall. and the head The simplest method of calculating mine water between the mine workings and the source recovery time is to assume no reduction in water aquifer starts to decrease. connections. term pumping rates for the mine. for discus- variable. K. R. but inflows are closely linked to aquifers may be readily available in the case of rainfall and will have a more dominant effect on the major aquifers. Coal Measures inflows water.1 ) and origi. Any deeper waters can have chloride values in excess seasonal variations in water inflow to a mine can of 200 000 mg 1-1. Figure 12 shows carried out by looking at the topography and an example of shallow mine workings recovery geology of the mining area under consideration . This approach yields a to decrease logarithmically with depth. The qua- the level of the base of the aquifer supplying the lity of shallow workings water can be very water. Schematic diagram of a general conceptual model for water inflow to mine workings.1 ) from Coal Measures aquifers. nate from minor aquifers. The exact course of water may not Data on the piezometric head in the source be known. inflows as the water level in the mine recovers Water quality from these sources is very (see Sherwood 1997. Where significant volumes of water enter shallow workings. are known sions of this point). 10. Larger flows may occur from major sandstones both above and below workings and from faulted ground. This water will only gradual reduction in head difference between the affect the older shallow Type A workings but can source aquifers and the water level in mine gravitate to the deeper workings via mining workings. Coal Measures aquifer water and surface are generally small (<330m 3 day. This method of Figure 11 shows a simple conceptual model of modelling will only be accurate when all the inflow to Cronton Colliery in Lancashire based water entering a mine originates at very shallow on a small make (72m3 day-1 from Type A depth. Chloride levels. This can be other inflows have decreased. (Glover & Chamberlain be addressed by using an average of the long- 1976). In these cases a reduction in water flow is required to reflect the Shallow workings water. Shallow workings water will include shaft Coal measures inflows. either from the shafts or from day. but in many cases there are no water levels in the later stages of recovery when data and so have to be assumed. WHITWORTH 69 Fig. for example. variable. and a moderate flow (459 m3 the mine at depth. the Coal Measures. a simple linear recovery is likely to be highly inaccurate.

snowing annual fluctuations related to seasonal variations in rainfall. and estimating pre-mining water levels from water inflow between the mine inflow level and springs. 11.70 MINE WATER RECOVERY. . MONITORING AND MODELLING Fig. 12. Simplified conceptual model of mine water inflow to Cronton Colliery. Mine water recovery in snailow workings in YorKsmre. water courses and aquifer outcrops. A Once the maximum head of the inflow source has straight linear reduction in inflow could be used been established. several methods of reducing the but the evidence from East Fife suggests that an Fig. the maximum head level can be applied.

using Darcy's flow equation. and V is the velocity or specific discharge.165 X 10~3 mday.537 X 1(T ms" ) 3.021 X 1(T7 ms~') . The permeabilities for each for a theoretical zone of interaction 30m above block were generally similar (see Table 2) and below the boundary of the total extraction suggesting that in multiseam workings there is area of workings.82 X 1(T m day" (1. The boundary is based on the square root of the area of total extraction where K is the hydraulic conductivity. rv. 13).135 X 10-8 ms" 1 ) 4. and A is the A simple method for predicting the reduction of cross-sectional area of interaction. possibly related to strata can be calculated. the head difference divided by the length of flow path will also be a constant (Darcy's law): where K is the Hydraulic conductivity. i. Table 2.581 X lO^ms" 1 ) 4. As the mine water recovers the is the total area of interaction around the mine inflows to the workings that originate from the workings.55 X 10"33 mday"'1 (7.1 (7.e. R. Average Coal Measures permeabilities based Average permeability model on the interaction area around total extraction This method of predicting water inflow reduction areas were calculated for several mining blocks is based on an average permeability calculated in South Lancashire. Bickershaw-Golborn 1. using the mining interaction area and the average permeability. WHITWORTH wi 71 constant. K. Fig. and A to be square. The relationship between total water inflow Using the total Coal Measures water flow into from the Coal Measures strata and the area underlain the mine or mining block and the area of by workings for four mining units in South Lancashire. /i2 is the minimum hydraulic head (in the mine workings). an average permeability can against the approximate boundary length of the be calculated: total extraction area for four mining blocks in South Lancashire shows a general linear relation- ship (Fig. area based on a zone 30 m above and below the Plotting of the Coal Measures water inflows workings boundary. h\ is the maximum hydraulic head (in the source aquifer). therefore. Cronton 3. Ashton Green-Bold 6.612 X 1(T3 3 mday"11 (1. an average permeability of the Coal Measures exponential reduction in flow. The dip of the Coal Measures a general relationship between Coal Measures strata in each mining block is assumed to be inflow rate and total extraction area. Coal Measures are decreased in proportion to The water inflow to the mine from the Coal the area or boundary length of workings still to Measures is then calculated for various depths be flooded. Agecroft 8. Q is the multiplied by 4. iv. Permeabilities derived from mine water inflow and interaction areas for five mining blocks in Lancashire Mining block Average permeability 1. the working area is assumed flow from Coal Measures strata at closure. the water inflow during recovery was developed Using the flow at closure and the interaction by relating the inflow to the area of workings. The Coal Measures flow the area of the drainage. the cross-sectional area of interaction: Are a-related flow model where Q is the flow. Vis the velocity. interaction around the total abstraction mining.865 X 1(T8 8 ms"1 1 ) 2. Sutton Manor-Clockface 6. L is the length of the flow path. 13.92 X 1(T m day" (4. as in a pumping test. is entering a mine is equal to the flow velocity times more appropriate.

the controlling factor is not usually the area of working but the permeability of the mining connection or the shaft lining. However.72 MINE WATER RECOVERY. Figure 15 shows several also applied to the deeper Coal Measures inflows linear mine water recovery curves based on to give a comparison with area-related and varying combinations of residual void space for average permeability models. MONITORING AND MODELLING Logarithmic flow model The area-related and average permeability flow models are methods to reduce Coal Measures flow into the mine. These predictions can then mining units assessed in Lancashire. However. Results of mine water recovery modelling To date. Figure 14 shows the predicted recovery curves at Cronton. of the four major mine water discharges. In these cases a simple calculation of 20% for total extraction and 30% logarithmic scale was applied between the inflow forpillar-and-stall workings. site and flow rates of potential mine water discharges can be predicted to reasonable accuracy. reductions in the flow predictions in the mining area or average Coal Measures per- meability can be used once the water in the mine Fig. if the void- level and the maximum head in the shallow mine space estimation is wrong the recovery period workings or the aquifer. there has been no monitoring of the The primary aim of mine water recovery actual mine water recovery in a mine or mining modelling has been the prediction of the timing. Where there is no mine have the longest recovery. the Environment Agency for allowing publication of . 15. However. Recent boreholes have water level monitoring and recovery has to be confirmed that mine water levels are still at depth calculated from an estimated inflow and an in the Bickershaw-Parsonage block and that estimated void. water levels in the Agecroft area are close to space calculation. There is approximately 30 years difference between the shortest and the longest period. Agecroft be used to design and correctly place remediation was predicted to have mine water recovery near- schemes to either prevent or to treat the potential est to surface and Bickershaw-Parsonage to mine water discharge. based on a residual void-space carried out using mine water inflow data and mine plans alone. Where mine workings extend above the levels of the inflows from shallow workings or shafts. (including the simple linear trend) generally give very similar recovery curves except in the late Conclusions stages of recovery. 14. where there are now for various estimates of residual void volumes. using the different The modelling of mine water recovery can be inflow models. The author would like to thank the Coal Authority for Fig. A similar technique was could vary significantly. the principal problem is the void. With this information the timing. However. this should be coupled with detailed monitoring of the actual water levels especially in the later stages of recovery. The various flow models used surface. shallow mine workings an alternative method of flow reduction is required. Predictions of mine water recovery at Cronton permission to use the mine water monitoring data and obtained using five alternative inflow models. to achieve accurate predictions on timing. Variations in predicted mine water recovery has reached the level of the shallow inflow or the times at Cronton Colliery using the linear inflow model shaft aquifer. block where recovery modelling alone was used the site and flow rate of potential future surface to predict recovery. Linking mine water inflow data with void- space calculations and monitored recovery data will give a better estimate of the reduction in water flow during recovery and help in calculat- ing the residual void space in a mine. both total extraction and pillar-and-stall work- ings. For flows from shallow workings or from aquifers within shafts.

Predicting Mine National Coal Board. & YOUNGER. Technical Report.W. ADAMS. Subsidence Engineers 1995. Colder Urban Environment. Groundwater in the the North East Leicestershire Prospect. Ground Water. YOUNGER. burgh. groundwater rebound after coalfield closure. National Coal Board. 1997. 2001. 99-105. R. P. The NUTTALL. Fife.L. Hydrogeological Assessment of CHILTON. 32. Croxford and J. J. N. 21–27 Septem- approximation using the GRAM model. Proceedings of the XXVII BURKE. P. Cornwall.S. Edin- Board. Mining pumping from the Frances and Michael Collieries. Dannatt of IMC Consulting of flooded deep mine voids by test pumping at Engineers in researching the mining data and the the Deerplay (Lancashire) and Frances (Fife) development of the Excel spreadsheet models. (eds). P. J. UK. 89-97.33. 2002. 1982. (eds). Mine Water Hydro- model of rebound in South Crofty tin mine. R. modeling ground water rebound in abandoned deep Coalfield closure and environmental consequence mine systems.M. geology and Geochemistry. In: YOUNGER. R.L. R. Developments in Geotechnical (Precautions against Inrushes) Regulations 1956.J. Modelling 362-365. 198.A. The Coal and Other Mines 5-7 April 1982. (eds). S. London. In: YOUNGER. 1999. Water indicator of the mechanics of rock deformation quality systems in Coal Measure Formations. Quarterly ber 1997.L. 1986.P. University of Newcastle. I. P. 39. P. & JOHNSON.25–2.. 315-326. . Ground Water..L. Modelling Minewater Flow istry. A. BARBOUR. 33. Groundwater Congress of the International Association of rebound in the South Yorkshire Coalfield: a first Hydrogeologists.A. National Coal Board. 198. National Coal Board. Water Around a Coal Face.. SHERWOOD. 4-7 September 1995. Nottingham. Special Publication. Geological Society.H. A strategy for ROBINS.the case in south Nottinghamshire. YOUNGER.). Predicting the consequences of ceasing Handbook 1975. 1976. Mining Department. & YOUNGER.L. WHITWORTH. & SHERWOOD. Rotterdam 165–170. London. Induced changes in the ogy. P. ADAMS. In: British Coal. Proceedings of the Fifth National Hydrology Unpublished Branch Seminar 1978. & YOUNGER.M. & ROBINS. . R. P. N. (ed.M. N. Geological Society. In: BLACK. British Hydrological Society. H. Engineering No. P. 198. Katowice. C. 1956. K. (eds). J. Mining Department. & YOUNGER. 249-261. London. 2002. permeability of Coal Measures strata as an GLOVER. M. National Coal Symposium.L. W179. 1997. DUMPLETON. 18-24. Journal of Engineering Geology and Hydrogeol. author would also like to acknowledge the help of Integrated-hydraulic hydrogeochemical assessment S. R. Department of Civil BANKS. & ROBINS N.R.L. Special Publi- References cation. In: FARMER.R. Special Publi.S. & ROBINS. Geological Society.S. In: above a longwall face. S. 1975. Balkema. Pro- Associates. Mine Water Hydrogeology and Geochem- istry. Unpublished PhD Thesis. P. WHITWORTH 73 the results of the South Lancashire modelling. mine water filling model. 2000. Mining Department. A. on Strata Mechanics held in Newcastle Upon Tyne. R&D Working Party Report 1972. & YOUNGER. Volume 1: Problems. and Quality Changes After Coalfield Closure. & ADAMS.G.L. E. Proceedings of the Symposium from Coal Mining Activities.C. Symposium on Environmental Problems Resulting Strata Mechanics. P. National Coal Board. head-dependent Engineering. SHERWOOD. Mine Water Hydrogeology and Geochem. K. In: YOUNGER. J. A. 2. cesses and Management. 2002. & WALKER. A variable-volume. National Coal Board. National Coal Board.L. Amsterdam.S. Collierres. Department. D.C. 149-160. Environment Agency. 1972. 39.L. Design of Mine Layouts. et al. & CHAMBERLAIN. Elsevier. Water Rebound. A physically based P. 2001. ADAMS. cation. (eds). 1978.

.

Mine Water Hydrogeology and Geochemistry.e. experienced several inrushes of groundwater derived from the overlying Lower Magnesian Limestone. Analyses showed that the water was derived from From: YOUNGER. Anomalous. causing serious disruption to coal production. London.S. with only 80m depth of of the sea caused by working too close beneath cover to the base of the Permian.1 (371s. Subsequent decrease in panel width coupled with increased depth of cover to the Permian reduced the incidence of water problems.5-m thick Barnsley seam at a depth of 550–600 m. how to prevent or minimize the (sometimes Mine workings in close proximity to aquifers disastrous) inflow of water to mine workings. with implications for enhanced aquifer recharge. threatening situation. mechanized longwall coal face inter. or other water-bearing formations such as peat With expansion of the industry from the early (Younger & Adams 1999) can experience severe years of the nineteenth century. production from the Barnsley mine workings encountering undocumented seam at Wistow Mine commenced in 1983. Widespread drawdown and recovery effects due to intermittent groundwater abstraction from a nearby factory were observed. 75 Geological Society. Precautionary observation boreholes had been drilled above and around the margins of two proposed longwall panels prior to working the 2. when peak transmissivity increases of 1979% and storativity increases of 625% occurred. Nearly all these incidents have been due to the North Yorkshire. the first longwall faces at Wistow Mine in the Selby Coalfield. several inrushes the sea bed was responsible for the deaths of 27 of water occurred causing coal production to be miners at Workington. P. 75-88. i. However. intra-cycle. UK (e-mail: sdu@bgs. there has never been any quantitative investigation to determine the effects of mining on the hydraulic properties of the stratigraphically higher Permo-Triassic age Sherwood Sandstone. recovery–drawdown events were also observed during this phase and interpreted as indicating rapid mining-induced dilation and compression of fractures within the aquifer fabric. but is largely one of safety.1 (Younger & Adams 1999). inrushes of water disruption. during and after mining. However. & ROBINS. . (eds) 2002.1 ) but gradually reduced to 3197m3 of old workings was highlighted tragically in day. 0305-8719/027$ 15. In the Selby Coalfield. Special Publications. the greatest effects were noted during the closest approach by the second longwall panel. Post-mining storativity remained mostly unchanged.1 ) during the following few weeks 1973 at Lofthouse Colliery. The results are consistent with similar investigations carried out at relatively shallow depths (<220m) in USA coalfields. abstraction well yield and possible increased contaminant transport rates. the Selby study shows that mining at much greater depths still has a significant impact on shallow aquifers.ac. flooded old workings. when (Bigby & Oram 1988. Kingsley Dunham Centre. Effects of longwall mining in the Selby Coalfield on the piezometry and aquifer properties of the overlying Sherwood Sandstone STEPHEN DUMPLETON British Geological Survey. Initial flow rates reached 10368m3 day. as well as posing a potentially life- have claimed the lives of hundreds of miners. a modern. although an inundation During the first year. N. The resulting inrush of mining industry. which also caused some additional subsidence over the first panel. 198. This opportunistic study fills that gap. Standard aquifer pumping test analyses of the hydrographs allowed transmissivity and storativity to be determined before. Cumbria in 1837 halted. with only 80m depth of cover to the base of the Permian. The results showed apparently permanent post-mining transmissivity increases of up to 149% around the margins of the panels. Nottingham NG12 5GG.00 © The Geological Society of London 2002. a major aquifer of regional importance. North & Jeffrey 1997). Keyworth. and up to 234% directly over the first panel. Data loggers permitted continuous monitoring of the Sherwood Sandstone and Drift piezometric levels over a 2-year period. West Yorkshire.L. water and debris killed seven men (Calder 1973). The impact of deep mining on groundwater has sected a flooded shaft whose presence was been a problem from the earliest days of the coal hitherto unsuspected.uk) Abstract: In the UK. However. The terrible hazard (1201s.

Selby Long Reach being Mine (Dumpleton 1996. Peace. This abstraction forms Subsidence Engineer. for 133(1). Booth. As a control. For example. such as transmissivity and storativity.1 (100 gall min. 1). West Virginia. piezometers fitted with automatic data documented. Subsequent stone. extracting 2. around Bank House Farm (SE BOCM). in the USA. at the request of completed in an overlying sandstone aquifer to UK Coal Mining. overlying Vale of York 25-foot Drift.45-m thick Pittsburgh No. The Barnsley seam at a depth of around 550–600 m. North Yorkshire. Sandstone at Barlby sewage treatment works. the aquifer is wholly confined by the now encounters only 'nuisance' water. Current mining at Wistow this study. demonstrating down at Unitrition–BOCM. This was a 600 m) on the aquifer properties of the Sherwood unique opportunity to investigate how these Sandstone in the Selby Coalfield. 83. comm. Barlby Reach and the by a retreat longwall panel (H93's) at Wistow east-west. 1).5m of The river is tidal and about 50m wide. the British Geological Survey (BGS) was commissioned jointly by UK Coal Mining Ltd Site description (formerly RJB Mining (UK) Ltd) and Unitrition The site is located about 1 km to the NE of Selby. instances of deep coal mining production.76 S. manufacturers of vegetable oil and 623 330) and Unitrition-BOCM (SE 625 329) animal feedstuffs. while. at the loggers were installed above and around the Windsor Mine. 8 Coal at also installed in an Environment Agency depths of 52-117 m caused monitoring wells and monitoring borehole completed in the Sherwood surface springs to become dry (Straskraba et al. although smaller inflows occur. to assess the potential effects at (Fig. Unitrition-BOCM abstracted ground- dust-suppression sprays. in the first few days of operation showed marked ing. Under tity pumped from the entire mine. e.) for public. private water supplies. In one piezometer. 1999). At Saline County. hydraulic conductivity of up to 1900% of its Initial inspection of the piezometer hydrographs original value were measured. p. interruption or deterioration of the groundwater More recently. was subsequently extended for decline by up to 30m. including water Environment Agency Abstraction Licence No. Monitoring was planned for Coal caused water levels in piezometers a year in the first instance but. other piezometers ing from weekday pumping and weekend shut- showed no significant increase. aquifer properties might change with time and. it was thought that the cycles of pers. second only to the Chalk of cover to the Permian reduced the incidence of (Price 1996. planned to be 170m wide. Illinois. Both of these examples deal with drawdown-recovery could also be used as a mine workings at shallow depth. an essential part of their operations and any comm. and One concern to be addressed was the possible flowed into the workings along fractures effect of the mining on groundwater resources in associated with the first mining roof breaks that the overlying Permo-Triassic Sherwood Sand- propagated up to the Permian rocks. In the area considered in red from time to time. topography is flat and low-lying. a data logger was in the 1. location removed from the effects of mining ings at a depth of 122m in the 2 m thick Herrin subsidence (Fig. 2 km long. area in a general southeasterly direction. during min. investigation was purely to monitor water level 1994.g. The present conftinuing series of aquifer pumping tests study investigates the effect of coal mining by yielding information about aquifer properties longwall extraction at greater depths (550. industrial and water problems. Sherwood Sandstone. The River Ouse flows through the Selby the Unitrition-BOCM Selby. Northern Illinois University fluctuations.1 ) (S. generally 3-4 m . UK Coal Mining Ltd pers. and widespread drawdown and recovery result- tude were observed. longwall work. per. Background to the present study In 1996. DUMPLETON the overlying Lower Magnesian Limestone. is reported to be only water from two boreholes drilled into the around 655 m3 day. Although the the complexity of the situation (Booth et al. temporary increases of an order of magni. a 1994). in times of greater environ. particularly. J. The quan. deliberately introduced into the mine. This extensive. supply would have severe consequences on mental awareness. longwall workings margin of H93/s. International–BOCM Pauls Ltd (Unitrition– North Yorkshire. a further year to observe any effects due to manent post-mining increases in the aquifer working of the adjacent retreat panel H94/s. The panel was the dominant features in the vicinity of the site. major aquifer is of great changes in mining methods and increased depth importance in the UK. with the site (SE 625 329) due to proposed coal extraction north-south. However. whether or not there were changes that could be attributed to the effects of mining. having a deleterious effect on both surface water Prior to the commencement of coal pro- courses and groundwater resources have been duction.).). C. 1 km long.

1. measured at the nearby Selby Selby sheet. Extending northwards from Bank House Farm. 71 (Selby) pers. The cross-section (Fig. comm. This compares favourably with based on Selby No. but the area chiefly arable and grazing for cattle and sheep. respectively. There is no Selby Long Reaches. and Barlby sewage treatment works (STW) is This is based on: (i) nearby shallow borehole and situated about 200m north of another east–west well logs. to Annual rainfall during the period 1941–1970 which the H937s panel made a closest approach was around 600mm (Institute of Geological of 300m. The gene- Barlby village lies to the east of Barlby Reach ralized geological succession is given in Table 1. . Artificial levees have been built on supplied by the Environment Agency). memoir available for the Selby sheet. land use is agricultural.). and (ii) National Coal Board Selby No. East of Barlby Reach and north of and 1:10560 Sheet No. EFFECTS OF MINING ON THE SHERWOOD SANDSTONE 77 Fig. SE 63 SW. the river banks to an elevation of 6-7 m OD. 3 borehole and the gene- annual rainfall of 480 and 528mm in 1996 and ralized vertical section of the BGS 1:50000 1997. Location plan. 3 exploration borehole (SE 61945 33323). reach beyond Barlby Reach. immediately south of the site is included in the while south of Selby Long Reach are the factory memoir for 1:10000 Sheets Nos 79 and 88 premises of Unitrition-BOCM and Rank Hovis. above Ordnance datum (OD) and prone to sewage works rain gauge (SE 634 313) (data flooding. (Goole and Doncaster) (Gaunt 1994). a prominent flood defence embankment of similar Geology elevation has been constructed from colliery spoil to offer further protection from flooding The geology of the area is covered by the British exacerbated by mining subsidence (S. Peace Geological Survey 1:50000 Sheet No. 2) is largely Sciences 1982).

un-named sandstone 7. Anhydrite. Horizontal and vertical scales are equal.78 S. the 259m of Measures proved southern limit of the H93's panel.A' (Fig. The 2. 16 Permo-Triassic Middle Marl Halite and marl.34. Anhydrite 5 (Edlington Formation) with marl beds.50 at borehole 429.44 at 535. occur in the Selby area. 3 borehole. with some peat Sherwood Sandstone Group Mainly red sandstones 175 with some marl bands Upper Marl (Roxby Formation) Marl. 16 Lower Magnesian Limestone Limestone.44 m Fig. The only below the base of the Permian consist of a cyclic sandstones of note are the Woolley Edge Rock.6 Quaternary Vale of York 25-foot Drift Interbedded sands and 10. Barnsley seam 2. soft 3 Carboniferous Coal Measures (Westphalian) Mainly mudstones. probably totalling some 750 m and at least 15 coals. marl 20 and gypsum 16 Upper Magnesian Limestone Limestone. Simplified geological cross-section along the line A . dolomitic.48 at 561. DUMPLETON Table 1.50.5-m thick Barnsley in thickness. Woolley Selby No. Marl 23 (Brotherton Formation) with anhydrite and gypsum. Lower Marl Mudstone with thin 1 limestone at base Basal Sands Sandstone. seat earths 19. 2.50 m thick. 3 Edge Rock 19. dolomitic. In Selby seam lies at a depth of around 550m at the No. and an unnamed sandstone 7.7–16. Generalized geological succession in the vicinity of the site Formation Typical lithologies Thickness in site vicinity (m) Holocene Made ground Alluvium Sand. siltstones 259 proved in and thin coals.2 clays. . 1). clay and mud up to 42 . succession of siltstones. mudstones. with 74 (Cadeby Formation) anhydrite in upper part.44 Coal Measures.1 -4.

industrial and private water drite. the anhydrite in the Middle Marl is a component. the dip of the base of the Permian in borehole. based on Selby No. 2. respectively.e. Above this horizon. aquicludes. where borehole and well records otherwise the sequence in the study area is indicate an interbedded sequence of sands and dominated by siltstones. Geological structure The Lower Magnesian Limestone (Cadeby Considerable detailed information is known of Formation) and Upper Magnesian Limestone the geological structure of the Coal Measures (Brotherton Formation) both form locally as a result of both extensive coal exploration important aquifers. The structure of the Permo-Triassic rocks is Permeability in the Sherwood Sandstone is nearly as well known as the Coal Measures. Gaunt (1994).6 Ml day"1) can horizon) dip in an easterly direction at around 1 be obtained. intergranular. This forms a local minor aquifer at or near outcrop appears to be broadly reflected at the Unitrition. with thin 60m northerly. The Common faults are present in this formation. In also a significant secondary fracture flow particular. In the area of of 70 m day"l have been measured. respectively. The dips of supplies. clays. the latter separ- These faults form the natural boundaries of a ating the Upper Magnesian Limestone from the block of reserves that was exploited from Wistow Sherwood Sandstone. ratios may be in the range of 1. BOCM site. (about 24km WSW of Unitrition-BOCM). Horizontal to vertical permeability prominent seismic reflector. i. The Vale of York 25-foot Drift (Quaternary) overlies the Sherwood Sandstone at the site. and included the H937s and H94's panels. especially in the basal 40 m. and peat and peaty silt overlie the and locally incohesive sandstone 0-30m thick. the Coal Measures (at the Barnsley seam yields of more than 301s"1 (2. and only the core sample studies from elsewhere in the Vale of . and it is in these areas that Fault is indicated on the cross-section in Fig. It is not clear from the seismic inter- Unconformably overlying the Coal Measures pretation whether the Barlby and Whitemoor are rocks of Permian and Triassic age. it is confined by the the fault planes appear to decrease in the Mercia Mudstone to the east and partly confined uppermost Coal Measures and the lower part of by drift. The Sherwood Sandstone Group consists In the Barnsley horizon the Whitemoor Common mainly of a brownish-red. seismic resolution is poor.to medium- Fault and the Barlby Fault. the former between the Upper and WSW. In the Coal Measures the Woolley Edge Rock sion. vicinity of the Unitrition-BOCM site. 25-foot Drift and occur over the entire site. terbedded. due mainly primary. and borehole interest. Lower Magnesian Limestones. although significant development is restricted by the limited thickness and recharge area. known. Infiltration occurs mainly where the drift the Permo-Triassic. EFFECTS OF MINING ON THE SHERWOOD SANDSTONE 79 thick. It forms a major these faults continue up into the Permo-Triassic aquifer of regional extent and of great import- rocks. 3 Generally. with individual units showing considerable earths likely to be of low permeability. In the underground mine workings. mudstones and seat- clays. In the Vale of York. These sediments represent the infilling of the Hydrogeology former 'Lake Humber' . may be regarded as typical for the the vicinity of the site is 1 in 34 to azimuth 076°.5 (based on however. which can form an effective aquifer. in- downthrows of 75– 100m southeasterly and 30. The Whitemoor Common is clay-free or absent. above the broad structure of the Sherwood Sandstone is Barnsley seam roof. The Middle Marl (Edlington Formation) and There are two main faults in the study area: (i) Upper Marl (Roxby Formation) both act as the Whitemoor Common Fault striking ENE. Alluvial The Basal (Permian) Sands consist of a friable sands. both overlain and underlain by sands.a glacially dammed The following summary is mainly based on the lake that existed in Devensian times in the Institute of Geological Sciences (1982) and southern part of what is now the Vale of York. but there is to the extensive coal exploration in the area. variation in thickness and lateral extent. Mine. very similar to the Coal Measures. have (less commonly coarse-) grained sandstone. 129 and 24m. the main public supply boreholes are located. ance for public.2-2. Gaunt (1994) described a silt and clay subdivi. although with reduced throws. hydraulic conductivities summarized in Figs 1 and 2. This information is vicinity of major faults. at least as far as the Middle Marl anhy. British Coal seismic surveys layers and lenses of brownish-red and greenish- (lines 94-SRL-32 and 90-SRL-01) indicate that grey mudstone and siltstone. and (ii) the Barlby Fault striking E-W. sequence given in Table 1. fine. in 30. having been exploited for (including boreholes and seismic surveys) and numerous domestic and farm supplies.

1997). indicate that the Sherwood Sandstone has a In the present study area there is little information transmissivity range of between 50 and concerning the hydrogeological characteristics 400m2 day. Where clay is ponents of ground water flow may be expected. represent. . thin or absent. A value of 450m 2 day.005-0. pumping test data infiltration is achieved (Spears & Reeves 1975). Dumpleton Mining subsidence etal 1998). 3). of the Drift. concentrating storativity available in the Selby area. although the latter was considered ex- cessively high. A pumping particularly on the area around and adjacent to test at Cowick. Prior to working. likely to be affected. presence or absence of clay cover. DUMPLETON York). Although in places overlain by was tentatively attributed to modification of the thin clays.006.1 . The presence of low-permeability marl infiltration to the Drift varies according to the bands suggest that significant horizontal com. analogue method gave generally comparable Sands and gravels of the Vale of York 25-foot results. (SEH) methodology (National Coal Board 1975) 1996). mining subsidence predictions There are few values for Sherwood Sandstone were made for H93's and H94's. these deposits are generally regarded subsidence 'angle of draw' by the Permian eva- as unconfined and vulnerable to pollution due to porites. day. following extraction of earlier panels (Wistow ing unconfined and confined conditions. which was most different values of 0. Calculations were made: The latter is thought to be more representative of (a) using the Subsidence Engineers' Handbook the confined aquifer storage (Aspinwall & Co. gave two the Unitrition-BOCM site. In Fig. completed in a Drift sand boreholes in the Sherwood Sandstone gave horizon. 3.08 and 0. although with a slightly wider (by about Drift are recognized as a minor aquifer (Aspin. 1996). (Fig. minimum values of 0. and (b) using an analogue method based Southern Water (1997) assumed maximum and on observed subsidence at Bondgate Lane.80 S.1 was considered to be representative of the true aquifer transmissivity (Southern Water 1997. pumping tests in two piezometer in BH3. At the nearby Hazlewood Pre. Selby. the degree of method. Following the installation of the serves site (SE 632 332). 2 the probable maximum limit of a permeable and locally thin unsaturated zone. The latter wall & Co. the actual and depth setting (Dumpleton 1996. Calculated subsidence profiles for H93's and H94's using the Subsidence Engineers' Handbook method (National Coal Board 1975). The value probably lies between these two extremes. Fig. 90–100% of the potential In the general Selby area. subsidence line shown is based on the analogue Elsewhere in the Vale of York. a simple falling head test yielded a transmissivities in the range 4 0 – 1 3 0 0 m 2 hydraulic conductivity of 6. 130 m) zone of influence at the surface.03 and 0.2m day-1. At the Hazlewood Preserves site.02-0. to the south of the area.2 X 10.1 . As the H73's and H90's) in a broadly similar geological aquifer was suspected to be leaky.0005.

95 44.52 7. elevations were determined by levelling carried Ferrari (1996) cites examples where working an out by H&H Surveys.00 7. However. therefore. The boreholes were drilled at weekend shut-down on Friday night or early 150mm diameter.).40-44.35 Drift (sand and gravel) Barlby SE 62585 35750 7. mining subsidence Unitrition-BOCM premises in a location where might have been directed along the fractures minimal subsidence effects were expected associated with the fault. trition-BOCM site. Pumping at South Side borehole is normally in order to minimize clogging of the borehole continuous from early Monday morning until the walls with fines. If the fault plane had been parallel. about dips in the opposite direction to the angle of 27 and 22 m. Subsequent to working of H93's and H94's. 1 (Olympia) out by Soil Mechanics Ltd in July 1996 using a and No. Each piezometer Coalfield (S. was fitted with an automatic data logger to record The effect of the Whitemoor Common Fault depth to water at hourly intervals.35 44.00 Sherwood Sandstone BH3 SE 62174 33601 2. in Figs 1 and 2 the fault in the Barnsley seam BH1 was situated outside the margins of the passes beneath and just to the north of the Uni. comm. but within the zone of subsidence.27 9. Summary details of monitoring boreholes Borehole National Surface Datum for Total depth (m) Screen interval Formation Grid Ref.85-8. at Unitrition-BOCM There are three boreholes on the site abstracting groundwater from the Sherwood Sandstone Piezometer installation under Environment Agency Abstraction Licence Drilling and piezometer installation was carried No. pump- internal diameter proprietary plastic tubing with ing is by airlift pump and is used intermittently to 1 m long screens set in a gravel pack and isolated top up a nearby header tank (Table 3).85 Drift (sand) BH4 SE 62854 32837 3.58 11. or line. Similar effects subsidence as mining proceeded during the have been observed elsewhere in the Selby course of the investigation. from the panel centre draw.76 3.45-44. Peace pers. who also carried out original panel of more than twice the amount periodic check levelling to determine ground predicted by the SEH method.35 10.45 Sherwood Sandstone BH5 SE 62854 32841 3. and a bentonite/cement grout. BH4 and BH5 were situated on the nearly so.70 43. H93/s panel.35-11. to the angle of draw. the fault plane BH2 and BH3 were directly over the panel. Borehole surface activation due to working of adjacent panels. elevation piezometers (depth below monitored (mOD) and data surface) loggers (m) (mOD) BH1 SE 62422 33093 3. The piezometers were 50mm Saturday morning. Data were on surface subsidence was not known at the time periodically downloaded to a laptop computer. possibly causing a (Fig.15 2. EFFECTS OF MINING ON THE SHERWOOD SANDSTONE 81 The observed subsidence at Bondgate Lane from the rest of the borehole by bentonite pellets probably included a degree of subsidence re. respectively. 1). Barnsley (contractors for adjacent panel caused subsidence above the UK Coal Mining Ltd). At Olympia borehole. 25 Unknown Sherwood BH Sandstone (control) .00-44.58 44. marked subsidence feature along the rockhead intersection of the fault plane (Dumpleton 1996).40 Sherwood Sandstone BH2 SE 62 17 133600 2. of making the subsidence predictions.76 3. although only No. As shown Table 2 summarizes the details of the boreholes. no such feature has been observed and.04 Approx.75 3. it Groundwater abstraction is likely that the fault has not had a significant influence on subsidence at the surface. 133(1)(A). Air flush was specified use. Table 2.70 43. 2 (South Side) boreholes are in current truck-mounted rotary rig.80 43.17 2.

82 S.14 m after working of H94's. hydrograph variations were identified that were underworked by the face on 14 November were attributed to features such as seasonal Fig. from a data Observed subsidence at the monitoring bore- logger installed in the South Side borehole as holes is shown in Fig. the latter part of the investigation. These took place from 27 November 1997 to 28 August times were checked independently by compari. and the control Barlby line conditions was very short.7 77.12m. Production from the adjacent panel. were recorded by Unitrition–BOCM staff. The pumping water level H93's was 0.77 Disused * These values based on rising main flow meter readings April . . times of pump 1996. 4.8-2. except in the case of BH2 where. This August 1996.2 No.September 1998. BH2 and BH3 BH. This can be com- During pumping shut-down.11 m for both panels (H94's did not commence until 30 August 1996. respectively (Fig. Observed subsidence at monitoring boreholes. 4. At the South Side borehole. 3 (Refinery) SE 62728 32845 approx. The face finished production on 29 March start-up and shut-down. 1996. monitoring sidence of about 0. was 0. about 120m Data analysis outside the H93's panel margin. 4 48. compares favourably with the calculated sub- due to an initially faulty data logger. Coal having little additional effect).4* (main abstraction borehole) No. 1 (Olympia) SE 62335 32828 approx. following working of H94's. H94's. 4 77.15-0.63m is typically about 40m below ground level. Abstraction boreholes at Unitrition-BOCM Borehole National Grid Surface elevation Depth Typical pumping rates Ref.31m. production from H93's started on 19 September In the Sherwood Sandstone piezometers. It can be seen that the part of a further BGS study (Dumpleton 1998. 1998. This increased to 0. 2 (South Side) SE 62800 32660 approx. BH1. The closest approach to BH2 (186m plan son with the hydrograph from BH4 and. Observed maximum sub- sidence due to H93's at BH1. and flowmeter readings 1997. to H93's and H94's of about 0. Monitoring of water levels commenced on 14 increasing to 0. maximum observed subsidence at BH2 due to Dumpleton et al 1998). so the time available for establishing base. 3. DUMPLETON Table 3.11 0. the water level pared with the calculated subsidence of BH2 due recovers to around 11 m below ground level. (m OD) (m) (Ml day.5 and 0. BH2 and BH4. 3).1 ) No.56m.11 1. during distance) occurred on 25 March 1998.

e. and storativity variations in BH1. 'best-fit' were obtained. Transmissivity Sandstone piezometers. Unitrition-BOCM draw. BH1 and BH2. and (b) smoothed by application of fluctuations. completed. Results down-recovery events and loading of the confined Sherwood Sandstone aquifer by diurnal Figure 5 shows typical smoothed drawdown– tides on the River Ouse. Hydrographs of piezometers com. pp. 81-84) to determine transmissivity and Changes in the aquifer properties are indicated storativity in each of the Sherwood Sandstone by the post-mining/pre-mining ratio. when peak values time series and compared with the timing and of 945 and 4568 m2 day-1 (increases of 466 and position of the mining activity to enable 1979%) were recorded in BH1 and BH2. to remove the effects of the tidal unsmoothed. The error bars indicate Barlby BH. The latter effects were time responses for BH1. 169). a 25-h moving mean to remove the tidal loading pleted in the Drift showed no effects due to effects. where t = time since pumping began and r = distance from Unitrition-BOCM pumping borehole. A three-point each of the pumping cycles indicated a leaky moving mean was applied to the 'best-fit' values aquifer response and were analysed manually to illustrate the changing trends more clearly. 12. res- conclusions to be drawn about the effects of the pectively. Values of transmissivity defined start–stop times and flow rate data. expressed piezometers. These parameters were plotted as approach by H93's and H94's. 5. Kruseman & de Ridder 1990. 'minimum drawdown fit' and pumping cycles were identified that had well. Both components were used when Unitrition–BOCM pumping and recovery. Unitrition-BOCM pumping cycle P9 (30/9/96-5/10/96): log-log plots of drawdown vs tlr2 of 'normal' smoothed (25-h moving mean) responses in BH4. Compare with Fig. Each cycle was cor. p. nor performing the manual matching to the Walton- due to tidal loading response. In addition. 100% indicating no change). using the Walton curve matching method Mean values.1 increased significantly method described in Kruseman & de Ridder in BH1 and BH2 during underworking or close (1990. . and and storativity were calculated in the usual way which could be identified in the Sherwood (Kruseman & de Ridder 1990). For each set of data a 'maximum A total of 84 individual Unitrition-BOCM drawdown fit'. Plots of drawdown versus time for the maximum and minimum fits. BH2 and BH4. BH4 were corrected for partial penetration Pre-mining mean transmissivity values in the effects using Weeks's modification of the Walton range 186–231 m2 day. the drawdown data for as a percentage (i. before. during and after mining are shown in Table 4. (Walton 1962. After working of both panels had been mining on the aquifer hydraulic properties. BH2 and BH4 rected for rainfall recharge by comparison with are shown in Figs 6–11. based on 'best-fit' results. EFFECTS OF MINING ON THE SHERWOOD SANDSTONE 83 recharge–discharge. Each filtered out as necessary using a 24 h moving piezometer hydrograph has two components: (a) mean. type curves. post-mining transmissivity values Fig.

7. Fig. BH2 transmissivity variations during period of investigation. DUMPLETON *ig. 8.84 S. . o. BH1 transmissivity variations during period of investigation. Fig. DM4 transmissivity vanations aunng period or investigation.

9. Fig. BH4 storativity variations during period of investigation. Fig. 11. BH1 storativity variations during period of investigation. BH2 storativity variations during period of investigation. . EFFECTS OF MINING ON THE SHERWOOD SANDSTONE 85 Fig. 10.

BH2 anomalous drawdown-recovery events (ADREs) within Unitrition–BOCM pumping cycle PI 05 (17/5/98-23/5/98). This is reflected the 2. The decrease seen responding to rapid mining-induced dilation and in BH4 is interpreted to be minimal. intermittent but heavy ground- rate within the main pumping cycle at Unitri.97 Xl0" 3 and 1. storativity appeared to be less affected by the mining activity than transmissivity with overall. small.1 (increases of 138-234%). they were of a magnitude transmissivity increases of up to 149% around and frequency that rendered it impossible to the margins of the panels. 12. storativity remained were interpreted as piezometric head fluctuations virtually unchanged (101%). Characterized by rapid fluctuations of The results (Table 4) show apparently drawdown and recovery within the main draw. 12). Possible reasons for slight tinuous monitoring of Sherwood Sandstone and departure from the type curves may include Drift piezometric levels over a 2-year period. Generally. How. associated with the times of storativity of the Sherwood Sandstone aquifer underworking by H93's and the proximity of to be determined before. during the Pittsburgh seam (1. and 234% directly over match the data points to the type curves (Fig. the first panel. during and after mining. Pre. Conclusions mining changes falling in the range 79–126%. undocumented adjustments in pumping Coincidentally. these variations Sherwood Sandstone piezometer hydrographs.44 X Elsworth (1999) in West Virginia. Compare with Fig. 5.04 X 1(T4 (224 and 625%) of around 215m. Automatic data loggers permitted con- minimum error bars. more dramatic departures were observed the hydrographs allowed transmissivity and only in BH2. possibly a Fig. H94's. were recorded. post. where 10" 3 -8. moderate.77X10" 4 . Post-mining Sherwood Sandstone These anomalous drawdown–recovery events storativity around the margins showed a decrease (ADRE) were classified on a visual basis as (79%) in BH4 and an increase (126%) in BH1.86 S. Similar effects were described by Liu & mining storativity values fell in the range 1. piezometers. Piezometers were installed above and around the Some pumping cycles proved more difficult to margins of two longwall panels prior to working fit to the type curves than others. USA. permanent post-mining Sherwood Sandstone down pumping cycle. fabric. In BH1 and BH2.5 m thick Barnsley seam at a depth of 550- by the range between the maximum and 600m. DUMPLETON remained higher than the original values at 257– compression of fractures within the aquifer 540m 2 day. . water abstraction from the nearby Unitrition- tion-BOCM and/or rapid changes in the general BOCM factory led to a widespread series of aquifer piezometric surface as a result of rainfall drawdown and recovery cycles observable in the recharge.7–1. Standard leaky aquifer pumping test analyses of ever. suggesting a general cause. marked or severe (Figs 8 and 11) and Directly over H93's in BH2. however.8m extraction) was closest approach by H94's. When they occurred. peak increases to worked by 183 m wide longwall panels at a depth 1. were noted in all three Sherwood Sandstone but little or no effect in the Drift aquifers.

J. Longwall working beneath water. Department of Trade and Industry. & MILLER. Nottingham (UK). Report on the causes of. International Mine Water Congress. Declining Groundwater Levels. Yorkshire. A. Inrush at Lofthouse Colliery Yorkshire. drawdown-recovery events (ADREs) were also observed during this phase and interpreted as indicating rapid mining-induced dilation and compression of fractures within the aquifer fabric. Coal International 26-28. and circumstances attending.W.EFFECTS OF MINING ON THE SHERWOOD SANDSTONE 87 function of incorrect pre-mining values due to insufficient time available to establish true baseline conditions..J. BOOTH. NR1711A. J. intra-cycle. particularly BH2. 1996. and the Potential Effects of Coal Mining at Unitrition International/ . J. London. North Yorkshire. CALDER. The investigation is the subject of a PhD thesis. DUMPLETON. in July 2001. the Selby study shows that mining at much greater depths has a more significant impact on shallow aquifers than had previously been recognized. The greatest effects were noted during closest approaches by the workings to the boreholes. In: Proceedings of the Fifth. During the closest approach by the second longwall panel. Problems of Ground Movement. Illinois. the inrush which occurred at Lofthouse Colliery. September 1994 IMWA & The University of Nottingham. the increased transmissivity and storativity in the 'surface zone' has imp- lications not only for enhanced aquifer recharge and abstraction well yield but also for increased contaminant transport rates in the event of any future pollution incident. Selby Region. Saline County. Within and immediately adjacent to the area affected by mining subsidence. 1996. Anomalous. References ASPINWALL & Co. However. D. Groundwater response to longwall mining. C. 1994. S. Unitrition International–BOCM Pauls Ltd and the Environment Agency. Appraisal of Groundwater Resources From the Sherwood Sandstone Aquifer. CURTISS. The results are broadly consistent with those from similar investigations carried out at relatively shallow depths (< 220 m) in the Illinois and Appalachian Coalfields.M. Report prepared for the National Rivers Authority.D. on 21 March 1973. submitted to the University of Sheffield. 71–81.S. The author gratefully acknowledges assistance from UK Coal Mining Ltd.N. & ORAM. peak transmissivity increases of 1979% and storativity increases of around 625% occurred. HMSO. USA. without which this investigation could not have taken place. Department of Civil and Structural Engineering. This paper is published here with the permission of the Director of the British Geological Survey (NERC). BIGBY. 1988.

1999. Selby. NATIONAL COAL BOARD.I. & DE RIDDER.C. STRASKRABA. Subsidence Engineers' Supplement to British Geological Survey Technical Handbook. The infiltration Preserves Ltd. Mining Technology. ution of carbonate. WE/98/16. Groundwater Temperature NORTH. Mine Infrastructure and Management. Environment Agency. 125–139. V. Geological Survey Technical Report. Wageningen. London. 1997. WALTON. Geological Magazine. T. Monitoring of Groundwater Report. Selby. Report. Levels in the Vicinity of Wistow Mine H931 sandH94' s Luxembourg. North Yorkshire: A Liu. SWINEHART. of the British Geological Survey. Mining Depart- Report. International Institute for Land YOUNGER. DUMPLETON. Analysis well and aquifer evaluation. J. Selby.P. 1998. Institute of Geological Sciences. M. 1997. 2nd edn. Selected analytical methods for KRUSEMAN. DUMPLETON.A. (Sheet 12). 49. & ELSWORTH. Pauls Ltd. 1990. D.. & JEFFREY. British PRICE. September 1994. D. WE/98/44. due to Wistow Mine Proposed H94's Panel. 1998. Hydro geological Map (UK). Determination of Subsidence 633-644.. Prediction of Monitoring at Unitrition International/BOCM Water Inflows into Coal Mines From Aquifers'. 1994. FERRARI. Predicting Mine Water Reclamation and Improvement. W179. London.A. Sheets 79 and 88 West Virginia. & ROCHELLE. National Coal Board. 1999.. European Commission Technical Coal Research. Produced for Dales Water Services Ltd as Aquifers of a Proposed Cooling System and agents for Hazlewood Preserves Ltd. Study of the impacts of a GAUNT. 2nd edn. Introducing Groundwater. British Geological water pressure fluctuation around an advancing Survey Technical Report. Survey Bulletin. S. Longwall Panels. C. 1996. Evaluation of pore Preliminary Investigation. European Commission. C. R. 1962. DUMPLETON. WE/99/19. USA. British rate into an aquifer determined from the dissol- Geological Survey Technical Report. S. PL. North Yorkshire. 22. IMWA & The University of of Southern Yorkshire. ment. & 78(898). Groundwater Recharge Scheme at Hazlewood SPEARS. 97/7/ Effects on the Sherwood Sandstone and Drift 1771. International Mine Water Congress. R&D Technical Netherlands. 171-176. Geology of the Country Around longwall coal mining operation on surface and Goole. BOSWORTH. British Geologi. & ADAMS. G.R. DUMPLETON BOCM Pauls Ltd. SOUTHERN WATER.W. North Yorkshire. EUR 15198 EN.A. 112.D. Selby. & REEVES. .J. cal Survey Technical Note. 1975. S. The case for continuing coal 585-591. W.C. WE/96/13. The Rebound. N. mining subsidence research. Publication 47. M. D. Final DUMPLETON. WE/96/13. 1997. Doncaster and the Isle ofAxholme. London. Test Pumping Report. longwall face. R. J. Illinois State Water and Evaluation of Pumping Test Data. 1996. SJ. Chapman & Hall. Nottingham Institute of Geological Sciences. FRANK. Memoir ground water resources at the Windsor Mine. Advances in Water Resources.88 S. North Yorkshire. In: Proceedings of the Fifth (England and Wales). G. W. 1994. scale 1:100000 Nottingham. S. M. BAKER.

. south. This model has already been applied to the simulation of groundwater rebound in several UK coalfields. Mine Water Hydrogeology and Geochemistry. 1997) and also prepared by the mining company with a consequent contraction of the Cornish tin (South Crofty pic 1998a. or (ii) the zone was never worked. the price of tin fell dramatically on world markets. 198. Newcastle Upon Tyne. and also from nearby disused and flooded workings. P. NE1 7RU. located in Cornwall. the hydrogeological characteristics of coal mines differ considerably from the South Crofty mine. 0305-8719/027$ 15. In the event. & ROBINS. Local professional opinion favoured interpretation (i). resulting in a marked steepening of the rebound curve during the later stages of rebound. 1999). as much as a year in advance of most other predictions. YOUNGER Department of Civil Engineering. because of bad experiences at the nearby Wheal Jane tin mine. These inflows usually originated from fault zones in the rock. N. During the 1990s the mine continued to commissioned by the Environment Agency. A retrospective simulation assuming very low specific yield in the 'blank interval' confirms that the hydrodynamics were otherwise successfully simulated using the SHETRAN/VSS-NET code. During several centuries of operation. The model was then used to predict the dates when surface discharges could be expected to commence. b). and closure was announced in were eager to obtain rebound times from a 1997.00 © The Geological Society of London 2002. has raised questions over possible environmental consequences. A physically based model of rebound in South Crofty tin mine. 89-97. In October 1985 Agency (Knight Piesold & Partners 1994. UK (e-mail:r. The quality of water draining from these shallow workings is generally good and no treatment is required. which could be interpreted in two possible ways: (i) the zone was worked in the mid-nineteenth century but the plans were lost.S.uk) Abstract: The recent closure of the South Crofty tin mine.adams@nclac. South Crofty mine. One of the main limitations of the predictions was a 'blank' depth interval in the mine plan records. 1996. who operate at a loss. However. (eds) 2002. and the prediction scenarios considered most probable were based on the assumption that mining-related specific yield values would be similar in this 'blank interval' to those applicable in better-mapped intervals below. The possible impacts of the closure of South west England. In order to address this problem and make predictions of the timing and volume of the discharge. with surface discharge commencing in November 2000. University of Newcastle. the last working mine of this type in Europe. (ii) above appears to have been the true case. and were found to be mainly head-dependent. with production switching to tin ore reports commissioned by the Environment from the mid-nineteenth century. In this mine. 89 Geological Society. The inflows to the mine during its operation were mapped and quantified. Older workings near the ground surface were dewatered by a series of adits that discharge into nearby rivers and streams.L. the mine was dewatered by a series of pumps located at different levels in the mine. began extracting copper ore in the Crofty have been well documented in a series of early 1600s. However. Special Publications. the country rocks comprise granite and metamorphic slates. the UK Environment Agency were concerned about the quality and quantity of water which was expected to discharge from the deeper workings when groundwater rebound was completed. Figure 1 shows a void-filling calculations (Adams & Younger plan of the modern mine and its environs. This case study was industry. which surround the modern mine. Most of the groundwater flow is therefore in the 'drives' and stopes from which tin ore was extracted. London. by which time South Crofty was the last physically based model rather than from simple remaining tin mine in Europe. strata that have negligible transmissivity and very low storativity. From: YOUNGER. Predictions of the water level during rebound are compared with the observed water levels in the main shaft which have been measured since the mine closed. Cornwall RUSSELL ADAMS & PAUL L. computer simulations using the SHETRAN/VSS-NET model have been carried out.

90 MODELLING REBOUND IN SOUTH CROFTY TIN MINE Fig. and flow Directive (Directive 76/464 On Pollution Caused via a short roadway into the main adit. EQS ensure that the mine water would emerge from are based on the EC Dangerous Substance the circular. comm. showing the extent of the modern workings. may result in increased concentrations a higher elevation than the Dolcoath Adit. therefore. the adit system was slightly modified to Standards (EQS) limits for these metals. this adit is at however. some of (Environment Agency pers. by Certain Dangerous Substances Discharged The Environment Agency's concern about the into the Aquatic Environment. however the quality of the water decanting River from the Dolcoath (Deep) Adit originated from South Crofty has actually been of better from long-abandoned shallow workings and was quality than that of the Wheal Jane mine water generally of good quality (Knight Piesold & due to the lower acid-generating potential of the Patterns 1997). 1. Location map of South Crofty mine. The 'first flush' of mine water. into the Portreath Stream. 1999) in order which date back to the eighteenth century. that rebounding mine previously received pumped water from the mine water would reach the Dolcoath Adit before of generally acceptable quality). It was of copper and zinc in the Red River (which had anticipated. A second adit (Barncoose) also minerals in South Crofty (Knight Piesold & drains shallow workings above the modern mine Partners 1997). mine water discharging into the Red Crofty. principal shafts and drainage adits. Figure 1 shows the location of treatment of the discharge has been mooted the mine and the various adit systems. however. The Agency was anxious to prevent a following a rebound of mine water. At to comply with the Environmental Quality closure. and associated pollution risk associated with the South Crofty daughter directives). and some Barncoose Adit. The Environment Agency became involved in discharge was partly a consequence of the serious the contingency planning for closure due to pollution incident caused by the discharge of concern over the high risk of polluting discharges mine water from the nearby Wheal Jane tin mine affecting the surface water courses (in particular in January 1992 (Knight Piesold & Partners the Red River which flows close to the mine). 1995). brick-lined Roskear Shaft. . Before mine similar incident following closure of South closure.

up to 900 m below ground surface. from the data and diagrams in the various reports The total plan area of the workings (as (South Crofty pic 1998a). showing the The 195 fathom pumps were used to pump workings (principal veins worked in the modern water from the lower workings to the ground mine are shown by thin black lines) and the surface. Cross-section north–south through South croftymine at Cooks Kitchen Shaft. mining engineers mine is the metric mining datum. . the South Crofty were at the 470 fathom (1200m dewatered volume of the mine was examined AMD) level. Each level formed pumped from four major pumping stations from the base of the stopes (workings). using the Cooks Kitchen Shaft. adit level 2051 m of the inflow rate of water (water make) during AMD.83 m) from adit level in the named Estimating the volume of the dewatered workings shaft to the connection to the workings. The other was extremely important in order to make accurate datum used in more modern references to the rebound predictions. named according to its depth and the shaft used to Between the 225 fathom (1695m AMD) level haul its ore to the surface (e. higher levels was negligible due to engineering Fig. Figure 2 shows a cross-section north. with and adit level were old workings which were the number referring to the depth in-fathoms partly dewatered by the 195 fathom level pumps. In tradi. south through Cooks Kitchen Shaft. The most modern and deepest workings in the operation of the mine. which were the 420 fathom (1295 m AMD) level (the deepest) driven vertically to extract the tin ore. ADAMS & P. in order to assist in the implemented in the conceptual model) is around setting up of initial and boundary conditions in 2 km2. L. The water make from the 195 fathom and dewatered area of the modern mine. R. this datum are referred to subsequently as 'm An assessment of the historical record of AMD'. Elevations relative to volume was estimated at around 4.g. therefore. 2). Ground surface (at Cooks Kitchen Shaft) pumping was made in order to derive an estimate is approximately 2111m AMD. YOUNGER 91 Mine layout Mine hydrogeology Workings in South Crofty were divided into a During the operation of the mine. water was series of near-horizontal levels. Secondly. to the 195 fathom (1730m AMD) level (the tional Cornish mining terminology. 310 Cooks). 2. (1 fathom = 1.5 X 106 m3. located 2000 m with first-hand experience were consulted. The below sea level (see Fig. the conceptual model described below. the level is shallowest).

were analysed. extreme wet condi. The analyses also identified a lag of app- from the long-term rainfall record for the area. Pumped mine water totals at different levels Level (fathoms) (m AMD) 195 (1760) 340(1495) 380(1422) 420(1340) 3 -1 Average pumped total 1989-1997 ( m h ) 273 273 124 48 Average make 1989– 1997 (m3 -1 ) 0 151 76 48 work carried out in the early 1990s to divert water Calculation of infiltration into the mine. In equation (1) the factor 0. extreme dry condi. as the correlation coefficients mine under 'best case' (i. between rainfall and pumping rate at the 380 tions) and 'worst case' (i. with the deeper levels contributing assuming that surface runoff equalled 75% of net less water make. it is clear that the levels Values of infiltration were calculated on a between 225 and 340 fathom (1495–1700m daily time series for January 1989 . local storm drainage and sewer system. and peaks in the pumping rate. regression analyses were performed using the clude the recent record up to 1998. infiltration) and the volume of water pumped from the mine. The results indicated that the inflows to the to the model (see below). pumping totals at the 195 fathom level pumps tude of the extreme conditions was predicted (0. The magni. The current baseline tion in water pumped from the mine. The from the shallow workings into the adit system infiltration into the mine was calculated from the and away from the mine (Carnon Consolidated rainfall and evapotranspiration data using 1990). and E assuming that the water pumped at each level is the actual evapotranspiration. The correlation coefficients were much times at which flow into the Dolcoath Adit from higher between the weekly rainfall totals and the Roskear Shaft would commence.25 is derived by make (55%).e. The 1961–1990 average dependent and which inflows originated from annual rainfall at the gauge nearest the mine was rainfall infiltrating the shallow workings and 1106mm.57). For the purpose of . totals and pumped totals was examined in order to A long record (starting 1950) was obtained from assess: (a) which inflows to the mine were head- three local rain gauges. In order to assess the likely timing of shafts. in which South Crofty largest inflows originated from the 290 and 260 mine is situated. comm. roximately 1 month between peaks in the rainfall The two scenarios are described below. 1999). Owen pers. immediately below. The aim was to produce deeper pumps were head-dependent and not from an appraisal of the post-closure behaviour of the infiltrating rainfall.e.e. local surface watercourses.January AMD) were contributing the bulk of the water 1998. where / is the infiltration. located between the 195 and evapotranspiration exceeds rainfall: 225 fathom (1700–1730m AMD) levels was used as a reservoir to store the water before it was pumped from the 195 fathom level. and (b) the time taken for fluctuations rebound in the mine it was necessary to study a observed in the rainfall to be shown in the varia- long record of rainfall. and produce an 'envelope' of different small. A large flooded area of old workings. The Redruth conurbation. extended to in. equation (1). fathom and 420 fathom level pumps were very tions). P is the rainfall. from ings connected via a major fault zone to the where the rainfall will be discharged directly to modern mine (M. The majority of the inflows rainfall. The high runoff is a consequence of having large paved areas such as industrial estates and car parks in the vicinity of Meteorological data the mine. In an urban district like the Camborne- originated from the eastern side of the mine. The meteorological data recorded at stations near the mine were obtained and examined in order to: (i) assess the likely recharge into the mine. The (ii) to examine the relationship between rainfall relationship between net rainfall (i. From Table 1. All units are in migrated downwards to the pump station mm. weekly pumping totals from the individual Two extreme scenarios were investigated and pumping stations to determine these relation- 3 year periods of monthly data extracted for input ships. infiltration can be as low as 25% fathom level (1584–1667m AMD) and were of net rainfall. No infiltration is possible if actual called 'the Pool'. This is due to direct runoff into the thought to originate from higher flooded work. A series of climate period of 1961–1990. and Relationship of pumping data with rainfall.92 MODELLING REBOUND IN SOUTH CROFTY TIN MINE Table 1.

2000). Schematic diagram of modelled inflows and the pipe network included in the conceptual model. Intact rock (granite and slates) is assumed was carried out using the SHETRAN physically to be of low permeability and assigned a based modelling system (Ewen et al. Therefore. this . YOUNGER 93 Fig. The different porosity values tions: are indicated by different shades of grey in • The lateral extent of the mine is defined Fig. Where two levels have the same by the location of the modern workings colour (e.g. porosity = 0. The model had been previously applied to UK abandoned coal mines zones. porosity = 0. ADAMS & P. 340 and 260 fathoms). Areas where the mine plans indicate stoping have a much higher hydraulic con- A 3-D model such as SHETRAN requires large datasets to parameterize the model domain. (Younger & Adams 1999).1 . incorporating the VSS-NET component (Adams low porosity and hydraulic conductivity & Younger 1997). 4. R. 3. The horizontal lines between the vertical Overview of the SHETRAN/VSS-NET model columns indicate the lateral flow pathways The simulation of rebound in South Crofty mine between model elements. shown in Fig. On average. ductivity and a porosity calculated from the In the case of South Crofty.14-0. depicting flow rates and location (by mined level).1%). this study.65%). Figure 3 shows the locations The mine was dewatered at the time of and magnitudes of the different inflows included closure to at least the depth of the lowest in the final model. head-dependent inflows model were set up accordingly (see below). L. except through faults and fracture up of old mine workings. The VSS-NET component (K = 10. outside the dewatered area of can represent: (i) laminar flow in variably satu- rated three-dimensional (3-D) porous media the mine. the model domain void volume divided by the total volume of was selected based on the following assump. 2. as insufficient data were available for any (1295 m AMD). the inflows were taken as average from 195 (1730m AMD) to 445 fathoms values. defined by the level of the lowest slopes on the 445 fathom (1340m AMD) level. there is negligible groundwater (Parkin 1996). accounted for approximately 60% of the total and the bottom boundary of the model is inflow to the mine. detailed analysis. Figure 4 shows a cross-section east-west Hydrogeological modelling of rebound through the SHETRAN conceptual model.5 mday. the workings (K = 1 mday" 1 . The initial conditions in the the pumping data. and (ii) turbulent flow in open flow even from saturated old workings to neighbouring unsaturated dewatered work- roadways or shafts as a result of the flooding ing. from analysis of workings.

equation (1). • 'Wet' . 4. grouping has arisen because both layers period during which the model results could have similar porosities. be compared with the observed water levels • The head-dependent inflows will be routed measured in the mine during rebound. evapotranspiration data were available allowing according to a schematic diagram (South daily infiltration values to be calculated using Crofty pic 1998£) (see Fig. the historical rainfall data and actual (1422m AMD) level by a pipe network.infiltration values were calculated from the 1965-1968 rainfall series (3 years commencing July 1965). calculated from the historical pumping The layers varied in thickness between 5m (at rates for 1997 and the first 2 months of and just below adit level) with 0. . level in the mine rebounded above the inflow level.5-m cells and 1998. Inflows from rainfall and • 'Historical' . model compared with the dips taken in the Cooks Kitchen Shaft. forming the water level rebound predicted by the the top boundary condition.94 MODELLING REBOUND IN SOUTH CROFTY TIN MINE Fig. 4). Cross-section east-west through Cooks Kitchen Shaft (in conceptual model). and element at adit level in the model. described the Roskear and Cooks Kitchen shafts in above.the model was run from 1 surface runoff were applied to each grid February 1997 to 31 December 1999. A pumping well (in The model domain comprised 167 125 nr the model) was located at the correspond- finite-difference elements vertically discretized ing grid element. For this to Cooks Kitchen Shaft at the 380 fathom period. and the pumping rate was using 200 cells representing 10 layers (Fig. 348 m with 5-m cells. After this simulation was run. This net. These data Simulations were used to run the model from January The first simulation carried out by the model was 2000 onwards. Head-dependent inflows wettest and driest 3 year periods extracted from were reduced gradually from an initial the 1961 to 1998 monthly rainfall (described maximum value to zero when the water above). 3). These scenarios comprised the either direction. These simulations are listed below. The average annual rainfall a baseline simulation representing an 'historical' for this period was 1269mm year" 1 . the work also allowed turbulent flow between two extreme meteorological scenarios. were run.

The increase in the rate of rise of water level during time axis begins in May 1998. substantially mined (no mine plans clearly bearing workings producing the inflows. under the 'dry' scenario by the grey pumping of the mine ceased in March 1998. YOUNGER 95 • 'Dry' . commencing March 1990). The model fathoms). The rate of rebound of mine water in was 'back-fitted' to the rebound curve by adjusting the upper levels would also be underestimated if the storage parameter in the top layer. against the observed subsequently failed to reproduce the sharp water level dips (shown as black crosses). as there Adit. 182m 3 h . the model predictions are 2000. Previously it was assumed that this layer was were no data on the head elevations in the water. the maximum discharge into the March 1998) to establish steady-state conditions adit was predicted to be approximately 121s"1. The average annual Future scenarios rainfall for this period was 955 mm year"!. In general. reproduced. Again. L. At the indicating workings in this layer were available. 6 by the recent rainfall data became available. however. obtained initially around 20m.1 . the model from the model results. . start of rebound. After black line. which the head-dependent inflows were reduced to zero extends from the 225 fathom level to the Dolcoath too early during the rebound process. head-dependent inflows totalled the plans showed a 'blank space' here). 5. The simulation was run until the water Results levels in the Roskear and Cooks Kitchen shafts were predicted to reach adit level (2051 m Historical simulation AMD). however. began water levels recorded during 2000 for compari- to rise. From January 2000 onwards. ADAMS & P. Again.infiltration values were calculated total inflow to the mine (depending on the net from the 1990-1993 rainfall series (3 years rainfall). this was approximately 60% of the the rapid rise in water levels from the spring of Fig. as before this time 2000. and in the model. line and the black triangles show the observed water levels in the mine. The results of the 'wet' and 'dry' scenarios within 20 m of the observed water levels. the simulation was extended after more under the 'wet' scenario is shown in Fig. The predicted water level in the Roskear Shaft Later. and the are summarized in Table 2 with the observed general rate of increase of water level has been time of rebound for comparison. which resulted in discharges into the there were few measurements of water level Dolcoath Adit commencing on 8 November taken. Figure 5 shows the predicted water levels son. There is a slight underprediction of In order to ascertain whether the model was water levels during the summer of 1999 once the capable of reproducing the actual rebound curve a water level reached 1650m AMD and this could final simulation has recently been carried out be due to an overestimation of the storage (again using the 'wet' monthly time series of coefficient in this level of the mine (260 recharge values from January 2000). these data were used to run the model from January 2000 onwards. An estimate of the flow rate discharging The 'historic' simulation was initially run for the from the Roskear Shaft into the Dolcoath Adit final year of mining operations (the year ending was also made. the 'wet' monthly time series of recharge values were used in the model. R. the underprediction of the levels was in Cooks Kitchen Shaft (solid line). by pumping water from the Cooks Kitchen Shaft. Observed and modelled water levels in Cooks Kitchen Shaft from June 1998 to January 2000.

Predicted rebound times under different scenarios Rainfall scenario Days to start of discharge Discharge starts Month when water reaches 2031 m Wet 1362 November 2001 September 2001 Dry 1500 April 2002 February 2002 Actual 983 8 November 2000 18 October 2000 Back-fitted 992 17 November 2000 September 2000 2000 indicated that the actual storage in this layer The model has also incorporated head- was much smaller than anticipated (back-filling dependent inflows of water into the mine with waste rock could have occurred). it was References possible to back-fit the rebound curve to match ADAMS. Table 2. However. However. by groundwater rebound in abandoned mines using a reducing the porosity in the top layer. J. . dry recharge and 'back-fitted' scenarios. treatment. 1997.96 MODELLING REBOUND IN SOUTH CROFTY TIN MINE Fig. Wright. Owen (Mining Geologist formerly with South (225 fathoms to adit) levels of the mine. workings from adjacent flooded mines through the porosity of this layer was reduced by 27% and faulted and fractured zones. These were thought the new prediction is shown in Fig. Thirdly. 6. it is possible that some of the old workings were backfilled with waste rock. Until early 2000 (when the mine water would emerge on 18 November 2000. A second reason modelling work made use of a UNIX workstation may be the extremely high rainfall recorded in funded jointly by United Kingdom Nirex Limited and southwest England in late summer and autumn NERC grant GR3/E0009. The main reason for the overestimate of the time to rebound is thought to be an The authors would like to thank the following overestimate of the volume of voids in the upper individuals for their assistance with this research: M. The study also showed a method of distinguishing head-dependent from rainfall- Conclusions derived inflows by regressing historical pumping The conceptual model developed to simulate rates with rainfall data. the Crofty pic). menced in November 2000. Simulation of the predicted date within a few weeks. discharges into the adit com. 6 by the dotted to be active during the early stages of the water line. Howells of results of this study indicate the importance of the Environment Agency (South West Region) and accurate measurements of the storage coefficient B. The new model simulation predicted that level rise in the mine. physically based modelling approach. C. 2000. with the benefit of hindsight. Current monitoring will rebound in South Crofty tin mine has predicted determine if the water discharging into Dolcoath dates of mine water emergence in the Dolcoath Adit is of poor quality and requires some Adit. Therefore. Sansom (Knight Piesold & Partners Ltd). The in predictions of rebound times. In: VESELIC. & YOUNGER. under wet.L. much earlier than predicted. Marsden and T. simulations were initially run) predicted water levels in the main shaft were within 20 m of the observed levels. P. Modelled water levels in the Roskear Shaft from June 1999 until emergence of mine water in the Dolcoath Adit. R.

saturated Subsurface Modelling System for River Basins. 250-258. 353-362. Minewater Study. YOUNGER. Report to the KNIGHT. 2000. Camborne. YOUNGER 97 M. South Crofty plc. South West Area. L. Bled. PIESOLD & PARTNERS. Bodmin. Bodmin. Environment Appraisal and nical Report. Dolcoath Deep Adit. PIESOLD & PARTNERS. RE. South Crofty Mine Closure. SOUTH CROFTY PLC. ASCE Journal Decant Location. 1994. 1998b. Environment Agency.. 1996. Final report to National Rivers Internal Report. Provisional report to Consulting Services Ltd. Vol. G. PARKIN. KNIGHT. South Western Region. University of Newcastle Upon EWEN. 1997. L. PIESOLD & PARTNERS. Nuwater Strategic Closure Study. (eds) Proceedings of the Sixth Treatment Strategy. South Crofty Hydrol. Bodmin. R. 'Minewate rand the Environment' 1. Report to the National Rivers International Mine Water Association Congress. & YOUNGER. Camborne. J. Tyne.A. 1998a. RL. & OCONNELL. P. Exeter. G. Slovenia. Unpublished PhD Thesis. & NORTON. South Crofty 8–12 September 1997. Timing of rebound at South Crofty. KNIGHT. R. 1995. Department of Carnon Consolidated. 1996. Predicting Mine KNIGHT. Unpublished Report by BUCKLEY. 2. RJ. Final report to ADAMS. SHETRAN: a coupled surface/subsurface model. J. R. South Western Region. A Three-dimensional Variably- ogy Study. 1990. PIESOLD & PARTNERS. South Crofty Environment Agency. CARNON CONSOLIDATED. Exeter. ADAMS & P. 1999. 5. Mine Closure Contingency Plan. Modelling the Environment Agency. Authority. of Hydrologic Engineering. ling system for 3D water flow and sediment and Report on its Condition and the South Crofty Mine contaminant transport in river basins. Wheal Jane Water Rebound. Civil Engineering. R&D Tech- Minewater Study.. Authority. Newcastle Upon Tyne. 1999. South Crofty SOUTH CROFTY PLC. . Report to Environment Agency. PARKIN. Environment Agency. & ADAMS. W179.

.

An evaluation of the AMD threat has been carried out with particular regard to the risk posed to the Sherwood Sandstone aquifer. 198. but these mine systems are relatively need for intervention was almost immedi. (eds) 2002. by aggressive AMD. issues are remedial action using cost-effective This created an urgent and desperate need to clean-up such as reed bed technology.ac. UK Abstract: The strata within and above the South Nottinghamshire Coalfield dip gently towards the east. if nece. S40 1JF.L.uk) 2 British Geological Survey. with field observations Durham Coalfields are also manageable in terms used to validate simulations with a satisfactory of their size and data assimilation. at an advanced stage of the areas of multiseam workings currently being From: YOUNGER. WALKER3 1 British Geological Survey. A very large area has been exploited with complicated internal drainage systems dependent on the maintenance of existing pumping regimes. the modelling. the 1999). 99 Geological Society. As a consequence. More importantly. N. Mine Water Hydrogeology and Geochemistry.the case in south Nottinghamshire N. with some certainty for individual collieries and the more so when effective clean-up method- for small groups of interconnected mine work. but to the east the coalfield is concealed beneath Permo-Triassic strata. The key took place throughout the UK during the 1990s.00 © The Geological Society of London 2002. . and the daunting any given time following mine abandonment and wealth of data pertaining to it. NG12 5GG. for the most part. mine water rebound and the acid mine drainage (AMD) cycle. to materialize in the UK. Nottingham. UK 3 Scott Wilson Kirkpatrick & Company Limited. Chesterfield. OX 10 8BB. In fact. analysis and degree of confidence.S. The evaluation has been assisted by three-dimensional (3-D) visualization that has enabled lumping of plentiful mine abandonment data. Wallingford. small. the constraints of South Nottinghamshire Coalfield has yet to be data handling require that larger and more abandoned.Robins@bgs. There are many abandoned shallow workings in the western area where the coalfield is exposed. overshadow the the rate of mine water rebound can be predicted prospect of sophisticated predictive modelling. DUMPLETON2 & J. the deep ings (Younger 1998). London Special Publications. ROBINS1. which overlies the concealed part of the South Nottinghamshire Coalfield. the mine as the Coventry and Staffordshire Coalfields has closure programme was effected so quickly that yet to peak. Similarly.S. ologies are available. However. Rose Hill West. particularly when compared with the ately replaced by need for remediation (Younger South Nottinghamshire Coalfield. Oxfordshire. The coalfield has yet to suffer closure. The coalfield retains The coalfields of Scotland and South Wales several deep working pits and there are large are. tunnels or landfill sites from attack predict environmental consequence and. These studies indicate that the critical spill-over elevation is 41 m above Ordnance datum (aOD). Bayheath House. S. UK (e-mail: N. complicated coalfields be analysed using lumped The South Nottinghamshire Coalfield is the data and this may have a detrimental effect on the last remaining large AMD problem area yet confidence of the results. Maclean Building. and predictive runs using the University of Newcastle GRAM model. P. and that the aquifer will be at risk about 20 years after pumping ceases from the Coal Measures. Kingsley Dunham Centre. A wide range of investigation and analysis problems currently facing the Northumbrian and has now been carried out. and the understand the physical and chemical processes protection of geotechnical structures such as that accompany mine abandonment in order to foundations. 0305-8719/027$ 15. Keyworth. However. Coalfield closure and environmental consequence . 99-105. to trigger intervention. Wholesale abandonment of deep coal mining mine water rebound and the AMD cycle. Discharge from areas such ssary. & ROBINS. the magnitude of the South chemical quality of acid mine drainage (AMD) at Nottinghamshire Coalfield.

for the aquifer north of Nottingham indicate a Fig. Schematic hydraulic connections within the South Nottinghamshire Coalfield. flooding of dry riers in the coalfield are shown schematically in workings. 2. ROBINS ETAL. and an outline cross-section of the oxidized products of pyrite. discharge is generally favourable. creates poor quality. The quality of the pumped mine In this area not only is the shallow and surface waters to date has been adequate for discharge to aqueous environment at the exposed coalfield streams and rivers without treatment. containing minerals such as soluble Fig. but private agricultural and consequential environmental damage would be industrial use is also significant. 2.100 N. dewatered. just as abandoned workings are also dewatered to it is or has been in all the other UK coalfields. 1. S. Fig. Up-dip under threat from the emergence of AMD. Although the quality of mine dewatering interconnections and perceived hydraulic bar. . Outline cross-section of the South Nottinghamshire Coalfield. Much of the abstraction is used for the pumps could be switched off or if the public water supply. but intercept water percolating down from the the important Sherwood Sandstone aquifer above exposed coalfield to the west where there are the concealed part of the coalfield might also be many long-abandoned shallow workings. 1. the Sherwood Sandstone the size and complexity of this coalfield there is Group forms the principal regional aquifer res- an urgent need to evaluate abandonment ponsible for meeting potable demand of nearly processes to determine the economics of whether 400 Ml day"'. Hydrographs too great. The at risk. Despite In Nottinghamshire. coalfield in the east is shown in Fig. exposed coalfield in the west and the concealed metal-rich water or AMD.

Boreholes and shafts penetrating the base of 200m in unmined areas because of the poor the Permo-Triassic. However. the Mainsforth complex of worked areas separated by in situ. 3). At the centre of the 10m rise in the piezometric head in the limestone sequence is the Top Hard. It has generally been considered that the Permian Marl sequence overlying the Coal The South Nottinghamshire Coalfield Measures was sufficient hydraulic seal to prevent The Carboniferous strata within the coalfield have ingress of mine water contamination into the an easterly dip away from the Pennine anticline. which is heavily exploited properties of the remaining strata to create a for public supply. the southern part of the Durham Coalfield. some faulted blocks of interbedded sandstone may be isolated from any source of recharge. and once dewatered may remain largely dry. but there are several large channel deposits that are up to 8 m thick. which has been worked aquifer. the marl is not universally collectively dip gently eastwards beneath Permian present and in places the Triassic sandstones marls. which act as discrete groundwater bodies. exploitation and for sensitive surface water and The widespread extraction of coal in south wetland sites. and the overlying on top of. The of mine water polluting an aquifer in the UK. and pumping stations. many of which deterioration in water quality in the aquifer. Rather than upward move- and seat-earths. longwall Measures are covered by the Permian Magnesian panel workings. but Colliery. and ironstones. . Nottinghamshire has greatly enhanced the trans- There is already one case of upward migration missive properties of the Coal Measures. Nearly all the deep mining inflow of Magnesian Limestone water into the has taken place within a sequence approximately former colliery workings. 3. mudstones. The extent interconnectivity of sandstone horizons (Rae of workings in the Top Hard. or rest directly generally thickest in the east. However. collieries that were situated in stippled. The strata are extensively faulted and vertical hydraulic connections are common across otherwise weakly permeable mudrocks and coals. The predominantly argilla- ceaous strata form aquitards that isolate sub- ordinate sandstone horizons. roadways and galleries. In creation of open shafts. Aided by mining- the Permian outcrop. these same fractures allowed flow to in Westphalian A to the High Main seam in occur in the reverse direction rapidly creating a Westphalian B (Fig. CLOSURE AND CONSEQUENCE IN THE SOUTH NOTTINGHAM COALFIELD 101 slight downward trend in piezometric level over local synclinal areas were far wetter than those on recent years. is situated close to the feather edge of disturbed. One mine. The sandstones in this area generally constitute less than a quarter of the succession. inadequately sealed Coal Measures have been extensively exploited in boreholes and shafts may provide a hydraulic the South Nottinghamshire and Derbyshire connection across the marl (Fig. country rock. were linked underground through this seam. Once pumping had 350–500m thick between the Blackshale seam ceased. which are typically fine Sherwood Sandstone aquifer associated with grained. which are in places locally absent but are come into close proximity with. In addition. and these strata take place. 2). 2). goaf-filled. This was accompanied by a significant to exhaustion in many collieries. if any. Frost (1979) reported sustained over the natural state. coals vertical fracturing. the hydraulic conductivity subsidence fractures that locally propagate up of the whole sequence is now greatly enhanced into the aquifer. with grey siltstones. the Coal Measures strata (Bishop & Triassic Sherwood Sandstone aquifer (Fig. However. reported small areas of saline water in the tions of grey sandstones. The Coal Measures comprise alterna. natural water transport occurs beneath depths of about Fig. Little. overlying Sherwood Sandstone aquifer should The Millstone Grit Series lies below the Lower significant mine water rebound be allowed to and Middle Coal Measures. has altered the hydrogeological Limestone aquifer. The Rushton 1993). and there is a concern from both the anticlinal ridges because groundwater tended to Environment Agency and local environmental migrate down-dip in the Coal Measures (Downing groups for the sustainability of current resource etal 1970). and there are numerous induced subsidence. Coal together with collapsed. Finch (1979) Coalfields. Blackshale and Kilburn is 1978). The Coal Measures Group constitute a complex multilayered aquifer.

such as mine pumping stations and reported carried out by analogy to abandoned coalfields wet areas where mine water makes have been elsewhere and by the simple estimation tech. high porosity goaf and fractured Three key horizons were selected for digitization: rock areas. he concluded that (2000). particularly high. even using fill and mine water would rise towards the Sher. This is because the coalfield is large and acceptable given the already coarse nature of the contains a complex interlinked system of mine water balance estimation (Table 1). as well as a number of Measures (Fig. employed. Mine abandonment plans were more diffi- to the dewatering volume currently taking place. It can be visualized as a set of interlinked and visualization package for data assimilation. the ponds would slowly (Dumpleton et al 2001). probably a roadway or interconnecting and other flow paths may open up allowing mine adit between collieries. Neither are easy to establish with any the Top Hard seam was adopted as the central degree of certainty. Groundwater Rebound in downward transport of groundwater from the Abandoned Mineworkings. It was the latter that was adopted for and working dewatered collieries. ment of mine water. A key to the evaluation is the identifi. As a general rule the sulphur content is the pockets of salinity derived from downward greatest wherever marine bands occur. Features that were incorporated niques such as those advocated by Younger into the model included: . The quality of the generated AMD depends on important coal seam. The University of New- Water make in the concealed coalfield from castle program GRAM. once the concealed coalfield is flooded on planar recovery of levels in discrete ponds. be incorporated into the model. process can The present-day dewatering scheme (as at be formalized by adopting a three-dimensional January 2001) protects the remaining working (3-D) geographical information system (GIS) pits. the surface water to migrate towards and into the Sherwood and groundwater modelling package SHETRAN Sandstone aquifer. workings penetrating a variety of different coal The visualization of the southern portion of seams. the volume of recharge equates model. Mine water risk evaluation usually difficult and often subjective. (Sherwood & Younger 1997). Besides. flooded. however. This model is based However. this 3-D package. the High Main the upper the pyritic sulphur content of the strata being seam. ROBINS ETAL. has been successfully Permian and Triassic strata is small to negligible deployed to predict mine water arisings at the compared with down-dip flow of water recharged surface in a number of different colliery settings directly to the exposed part of the coalfield. There are a number of commercial packages The shallow ponds are recharged locally and flow available including EARTHVISION and VUL- down-dip to be intercepted by pumping stations CAN. as well as emerge at available data could usefully and cost-effectively surface in the exposed coalfield. the range of different types space that needs to be filled to saturate the system of plan and the accuracy associated with them. fractures critical lip. 3). the South Nottinghamshire Coalfield concen- cation of the main recharge process to the system trated on bringing together the mapping and in order to quantify the volume of incoming water. and the Kilburn the lower seam of interest. and the base of the Permo-Triassic sequence is each overflowing into the next lower one at a subject to a reversal in head gradient. cult to deal with because of the volume of data that A second important parameter is the actual void needed to be digitized. it was found that not all the wood Sandstone aquifer. S. The distribution of sulphur in the South Of the other 30 or so seams that have been worked Nottinghamshire Coalfield has not been mapped in the area. data were selected for critical areas and quality prediction can only sensibly be only. and some of these will in. there are many shafts has been used to evaluate mine water rebound and boreholes penetrating through to the Coal (Adams & Younger 2000). of mine voids. However. other packages that have also been successfully evitably contain imperfect hydraulic seals. borehole geological data beneath a digital terrain In broad terms. If the pumps analysis of the South Nottinghamshire Coalfield were to be switched off. and partially overflowing underground ponds. One of the basic problems in the larger-scale evaluations is the bringing together of diverse sets of data to create a conceptual model.102 N. which were losing to have been made to address the problem of AMD the sandstone aquifer. This. More recently. percolation of pumped mine water that was A number of different modelling approaches discharged to surface waters. breakthrough prediction. and data lumping Quantification of this risk and prediction of was an essential part of the model design. It was both breakout time after pumping ceases and of considered that the implied approximation of water quality in the future is both difficult and mine void data caused by data lumping was fraught.

After the initial flooding of the mine complex. therefore. It is. 4). on the Coal Measures in the concealed coalfield will attain an elevation approaching 41m aOD. that as near steady state is approached the head selected shafts and boreholes. CLOSURE AND CONSEQUENCE IN THE SOUTH NOTTINGHAM COALFIELD 103 Table 1. It is assumed seam contours and spot levels. of which stream flow Groundwater abstraction 0. This becomes the key control to mine concealed coalfield. workings area are shallow and old shafts exist at to between 40 and > 100m aOD along the wes- the former Radford and Newcastle collieries. likely to be small. perhaps locally constrained in relatively undis- tified a set of 'hotspots' at which AMD arisings turbed areas to a slightly lower head. 4. and it upward groundwater (mine water) flux from the is this elevation that prescribes the risk of mine Coal Measures and where the aquifer is under- water rising into the Sherwood Sandstone aquifer mined. Schematic cross-section showing likely surface discharge and groundwater fluxes to and from the Sherwood Sandstone aquifer at the critical mine water recovery level of 41 m aOD. Virtual flooding of the visualization model iden. areas of goaf. a critical low-elevation area occurs near the concealed coalfield ranges from only 20m Radford (SK5498 4101). where the Top Hard aOD in the eastern down-dip part of the aquifer.5 separation and distribution of sandstones suggests recharge may amount to 35 * The water balance assumes that the rivers gain as baseflow the difference between estimated recharge and mine water plus groundwater pumping. field and down-dip beneath the concealed coal- cross-measures drifts. principal in-seam mine roadways linking The head difference between the concealed coal- areas of goaf. that is potentially vulnerable to migration in the concealed part of the coalfield (Fig. . roadways and other effective pipe-flow voids. In the exposed nal head in the Sherwood Sandstone aquifer over coalfield. where there is both an water recovery levels in the whole coalfield. critical spill-over elevation to ground surface is the critical central and eastern area of the estimated to be 41m above Ordnance datum Sherwood Sandstone aquifer above the worked (aOD). of AMD. Outline water balance * for the catchment area of the south Nottinghamshire exposed coalfield (after Dumpleton et al 2001) Inputs Volume (Mm3 year [) Outputs Volume (Mm3 year *) River flow onto Not known River flow out of the 107 catchment area catchment (Rivers Amber and Erewash) Sewer water/main Small Mine water pumping 9 leakage Effective precipitation 77. The regio- may potentially cause problems. The tern feather edge of the aquifer. Borehole-specific capacites in the Sherwood small volume laminar flow will prevail through Sandstone aquifer range typically between 1 and Fig. therefore. field is.

for the . Deteriorating seals in essentially crude. The lapsed time between cessation of pumping or shut-down of selected pumps to AMD discharge Conclusions at the critical elevation of 41 m aOD ranges from Mine water recovery in most of the UK coalfields only 20 years if the system is modelled as a is at an advanced stage and requires remediation single pond to over 50 years if it is considered as rather than prevention. elevation of mine water discharge is at 41 m aOD given the small difference in head between the in the main. The identifies critical areas in the concealed coalfield.e. In the eastern below. there is an upward flux from the Measures.81s"1 m"1. given shut-down effect on low-flow quality. pumping ceases using GRAM and a variety of on the overlying Permo-Triassic strata have not different pumping scenarios. Calverton and Blidworth mine quately sealed against the rising head in the Coal workings. S. Drawdowns of this order model reveals areas of potential risk to the have obvious implications for potential upconing Sherwood Sandstone aquifer. for example through of all or just of selected pumps. model. as a worst case. taken to attain these critical elevations after Abdy-Brinsley. the base of the Sherwood Interrogation of the visualization model rapidly Sandstone generally lies below this level. drawdown of 16m. would produce a draw. However.1 and a storage AMD generated in the system is not known. Coal Measures. effort was concentrated on digitizing prevent pollution in the Sherwood Sandstone three key horizons into the 3-D model. this critical elevation in the eastern extremity of mian marls are locally thin or absent. albeit of variable quality. These were based on been fully investigated. High Hazles and Top Hard seams. but coefficient of 0. has been success- the dilution of sewage treatment effluents in some fully carried out using a combination of 3-D rivers and streams. water balance estimate mine shafts and exploration boreholes may also (Table 1). is 1. only commence when pumping ceases. Inflow was derived from a simple. but is suited to the large area of the South Nottingham- not likely to cause significant environmental shire Coalfield and the overwhelming volume of damage. Most of the coal- criticality is possible. The controlling of AMD leaking up from the base of the aquifer. Sherwood Sandstone aquifer may have more available for this area. The the presumption is that it will be net alkaline and 4-day specific capacity for this site derived from a aggressive during the initial phases of flooding. AMD percolating into the base of the mine plan data.005. This modelling combination was best possibly also by interception pumping. AMD arising at surface in visualization of the data and the University of the Newcastle and Radford areas of the exposed Newcastle GRAM mine water rebound estimation coalfield can be dealt with by remediation. 21s -l m -1 so that a high yielding public supply Directive. with vertical inflow to the concealed provide similar flow paths. ROBINS ETAL. handling and modelling is a manageable task. particularly in the High Main. part of the aquifer and the recovered Coal Measures water South Nottinghamshire Coalfield. underground interconnections so that data assem- bly.104 N. where the Per. typical for this aquifer. fully interconnected. The aquifer is underlined in the EC Water Framework remainder of the mine plan data (i. This model indicates that occurring in a few coalfields. coalfield considered to be relatively insignificant. been worked in the South Nottinghamshire guard the quality of the aquifer.e. The South Nottinghamshire Coalfield is still Environmental consequence pumped and the mine water rebound cycle will Loss of present-day good quality mine water dis. Test pumping data water supply could be maintained with the loss of for boreholes at Edwinstowe (SK 6267) indicate a this resource). of the Bestwood. or where the South Nottinghamshire Coalfield above much shafts and boreholes occur that may be inade. This ensures the prevention of mine borehole. the precise quality of the transmissivity of 1500m2 day. continuing easterly dip brings the piezometric These occur where worked seams approach the level in the Sherwood Sandstone Group below base of the Permo-Triassic strata. i. but upward flow of mine water. It is possible that such either a single 'pond' system or a system of effects may create new pathways for future two ponds. from about fields are relatively small and limited to few 20 years onwards. Mine dewatering is still two separate ponds. The stacking effects of multiple Dumpleton et al (2001) predicted the time seam workings. Predictive charges to surface waters may have an adverse modelling of mine water levels. created by a pumping rate of Interrogation of the VULCAN visualization 28 Is" 1 . yielding 401 s~ ] . part of the Coalfield. serious consequences such that a programme of Given that up to 30 different seams have scavenger pumping could be required to safe. water ingress to the aquifer (even if the public down of between 20 and 40 m. The need to Coalfield.

1-29. Nuclear and Renew- able Resources. LOVE- field pumping rates. Evaluation of the rate of decrease in required in order to satisfy the requirement of the the iron content of water pumped from a flooded EC Water Framework Directive and to retain shaft mine in County Durham. At that time mine water may percolate 307-319. P. Modelling Permo-Triassic strata because of increased frac. Mine Water and total iron concentrations in groundwaters flowing the Environment.H. 19-21. Analysis of the adjacent coalfields to the north RAE G. R. Predicting temporal changes in Mine Water Association Congress. Quarterly Journal be longer depending on the complexity of the flow of Engineering Geology and Hydrogeology. Mine water rebound in South of the former Radford and Newcastle collieries. 1970. & YOUNGER.. the Technical Report Department of Civil Engineering. The Hydro- vital input offered justification for the lumping of geology of the Trent River Basin. J. YOUNGER. PL. K.W. R. PL. Poland. Model of the Sherwood Sandstone Aquifer. R. upwards within the concealed coalfield to emerge FINCH. 47—69. University of Birmingham. 1979. T. P. Processes and Manage- and Coal Measures in this area. 2001. system. P. Abedeenshire. J. Although mine plan data are abundant. S. be determined other than by equating it to coal. IMWA & Universytet Slaski. from abandoned deep mines: a first approximation. Simulating Coalfield: pollution risks and popular resistance. London. WALKER. it field. & surface at an elevation of 41 m aOD in the vicinity MERRIN. J. The coarse nature of this LOCK.. & YOUNGER.R. Geological Sciences. LAND. N. PL. 40. Katowice.) Energy and the Environ- permission of the Director. prediction that the coalfield will overflow to DUMPLETON. Journal of groundwater quality sufficient for public supply. University of Birmingham for the National Rivers total potential recharge to the coalfield could not Authority. DOWNING. 101-111. Hydrology.A. In: ture density and the relative geometry of aquifer CHILTON. the Environment Agency and NERC. mine. 1979. 1998. In: Proceedings of the Seventh International YOUNGER.R. Central Water Planning Unit. PL.W. etal. 218-228. degree of hazard to groundwater in the overlying SHERWOOD.. PL. & RUSHTON. the Yorkshire Coal- attempted with any degree of detail. Nottinghamshire: risk evaluation using 3-D visuali- This will take at least 20 years to occur.E. Water Resource and generalized for the construction of the Study of the Nottinghamshire Sherwood Sandstone GRAM model. R. 44. & BRIDGE. which was used to predict various Aquifer System of Eastern England: Mathematical rebound scenarios. Interception or scavenger pumping may then be FROST. PhD Thesis.S. D. .J.O. Published by In: NICHOLSON. 1997. 2000. L.C. 1999. CLOSURE AND CONSEQUENCE IN THE SOUTH NOTTINGHAM COALFIELD 105 thinner and lesser worked seams) was lumped BISHOP. Coalfield abandonment: geo- The work described in this paper was jointly funded by chemical processes and hydrochemical products. groundwater rebound after coalfield closure. Institute of other data. 1978.R. Hydrogeological The main outcome from the modelling is the Report 5. 1993. 1. References YOUNGER. Macregor Science. ROBINS. groundwater rebound in a recently closed tin Bands and Banners. England. 165-170.M.A. K. Restless waters of the Durham ADAMS. The Further Development of Elec- trical Resistivity Techniques for Determining Water at the base of the Sherwood Sandstone aquifer.. is suspected that these areas offer a greater Reading. Journal of Contaminant Hydrology. ment: Geochemistry of Fossil.J. Quality. 34. Technical Note. but could sation and predictive modelling. ALLENDER. However. (ed. British Geological Survey. Groundwater Resources in the Coal- of the current investigation have not yet been fields of England and Wales. Balkema. 2000. ment. (eds) Groundwater in the Urban Environment: Problems.. Rottedam.

.

In the Hamburgo Basin. This analysis showed that a drainage tunnel with horizontal drains was the most effective method of draining significant areas of the slope to the required pore pressure targets. 2 demon- at least three alteration types: (1) potassic altera. (eds) 2002. . a major programme of interceptor alteration. The mine has been as shown in Fig. of northern Chile (Fig. to the west of the 700-1200m to the east of the mine. while andesite is found mainly in the In 1996. It is tion seen as pervasive biotite and feldspar. BENJAMIN MIRA2. Antofagasta. 501 Avda de la Minera. The main structural features operating since 1990 and currently 900 0001 of are NNW oriented. N. or unaltered porphyry. Fig. A hydrogeological investigation has been undertaken as part of the depressurization operation on the north wall. and quartz-sericite. strates the conceptual model of flow to the pit. To investigate effective drainage measures for the silicified porphyry. com) 2 Minera Escondida Ltda. The Escondida copper mine. allows enhanced flow along the strike of the struc- Younger rhyolites cover and truncate the por.L. CRISTIAN VALDIVIA2 & WARREN HITCHCOCK2 1 Water Management Consultants. Special Publications. 107-119. It has shown that there are two distinct hydrogeological units in the slope. but is modified due to overprinting from the pit through specific structural offsets. Chile PATRICK MCKELVEY1. Shrewsbury. UK (e-mail:pmckelvey @ watermc. owned and operated assemblages infilling the fracture networks. 107 Geological Society. The by Minera Escondida Limitada (MEL). and define boundaries of hard- copper are extracted per year from a 400 m deep ness. barrier to flow perpendicular to the zone but cene age. This recharge flow is alteration haloes can be generally applied to the partially blocked by the Panadero Zone but enters deposit. 198. 23 Swan Hill. phyry. grades. a groundwater model of the slope was constructed. 0305-8719/027$ 15. GEOFF BEALE1. P. 1). in the Panadero. The underlying silicified. levels on the east wall of the pit. London. known as the Escondida Porphyry. alluvial Panadero. Pore pressures in the silicified porphyry are higher than the altered material. A system of horizontal drains and vertical wells has been in operation to reduce pore pressures in the altered porphyry. discrete seepage faces formed at high south and west of the pit. SIMON MANSELL2. mineralization type. & ROBINS. is located zone of advanced argillic alteration reaches its 160 km SE of Antofagasta in the Atacama Desert greatest extent on the north and northeast walls. ture (Water Management Consultants 1999). alteration and copper open pit. and (3) advanced argillic Since 1996. (2) calculated that a maximum of 701 s"1 is flowing phyllic alteration comprising chlorite—sericite to the mine from the Hamburgo Basin. There are leakage points.S. An evaluation con- WMC 1999). The Panadero Fault Zone on the east of the The copper mineralization is associated with mine is one of these features (see Fig. has very low matrix permeability and groundwater flow is in steeply dipping NW-trending fracture systems. the altered and unaltered Escondida Porphyry. firmed that groundwater was flowing from the The porphyry copper system of concentric Hamburgo Basin to the pit. The hydrothermally altered Escondida Porphyry is clay rich and has relatively high matrix permeability. Depressurization of the north wall at the Escondida Copper Mine. 2). 2. Chile Abstract: Concurrent dewatering and slope depressurization operations have been underway at the Escondida open pit since 1996. It forms a a quartz-monzonite intrusive of Eocene-Oligo. or multiple volcanic and intrusive events. which is present as pervasive clay well drilling has been implemented to cut off From: YOUNGER. Mine Water Hydrogeology and Geochemistry.00 © The Geological Society of London 2002. ADAM TAYLOR1. In addition. water level monitoring and colluvial deposits of gravel and sands form a showed that groundwater levels in the rhyolites groundwater storage reservoir of up to 100m to the east of the pit had risen by 20-30 m from thickness (Water Management Consultants the pre-mining water level.

specific hydrogeological investigations were implemented to identify the most effective depressurization methods for that sector of the mine. Approxi- mately 100 boreholes have been constructed on the wall. geology (alteration pattern and lithology) and water levels. Field programme and database Figure 3 shows the layout of the north wall. On most boreholes. Initially. The pore pressures within the in-pit rocks are the result of slow dissipation of pore pressure within low permeability material rather than the recharge flow from Hamburgo. The rim of the wall is at 3100 m above sea level (m asl) while the base is at 2700 masl. Currently. depressurization programme carried out on the north wall between work is underway to reduce pore pressures within September 1999 and February 2000. density. Subsequently. with the locations of boreholes and horizontal drains. In association In addition. rhyolites on the margin of the pit. . information (airlift flows and penetration rate). there was a targeted investigation with this interception operation. single and multiple piezometers were constructed to measure pore pressures at specific intervals. neutron and acoustic ceptor system abstracts up to 1501s. single standpipe boreholes were constructed to gain basic information on the phreatic surface distribution. 1.108 NORTH WALL DEPRESSURIZATION. Therefore. resistivity.1 from the televiewer. the eastern rhyolites. Location plan. ESCONDIDA COPPER MINE the pit. In this Fig. Geotechnical studies have been carried out to optimize the stability of the slopes. These studies have shown that pore pressures need to be reduced on the north wall. a downhole flow from the basin and to reduce water levels in geophysical suite was run comprising caliper. The dataset for each borehole comprises drilling Fig. Conceptual model of the recharge mine. the east wall inter. 2. gamma.

P. McKELVEY ETAL 109

Fig. 3. North wall layout.

hydrogeological investigation the work pro- • There is a difference in heads between the
gramme consisted of: altered and silicified rock. Silicified rock
heads are generally up to 30 m higher than
• The construction of multiple piezometer
heads in the altered material.
groups - to investigate pore pressures
• There is a reduction in permeability
in the various lithological and alteration
between altered material and silicified
units throughout the slope;
rock. This is seen in the packer test results,
• The drilling of four 300 m deep holes with
which are displayed in Fig. 4. Altered
packer permeability testing - to assess the
material permeabilities range from 10-4 to
drainability of each hydrogeological unit.
10- 6 cms~ 1 . Unfractured silicified rock
Interpretation of the results of the north wall matrix permeabilities are measured in
investigation programme, together with the data the range 10~ 7 -10~ 8 cms~ 1 . However,
from existing boreholes, has shown that (Water 10~ 8 cms~ 1 is about as low a permeability
Management Consultants 2000): as can be measured using packer test
equipment, and the matrix permeability of
• There is a multilayered rhyolite at the top some sections of the unaltered material is
of the north wall. Groundwater heads in expected to be lower.
the rhyolite are up to 160m below ground • These permeability values suggest that the
level (m bgl). Potential recharge flows to altered material can be globally drained,
this unit are partially blocked by the but that the silicified rock can only be
Panadero Fault Zone or have been cut off drained if a sufficient number of transmis-
by the interceptor wellfield. sive structures are intercepted.
• Below the crest of the slope, the geology
changes to Escondida Porphyry. The Approximately 100 drains have been drilled
thickness of argillic altered material is into the north wall. Data available for these drains
around 250m near the top of the slope. include position, orientation, length and weekly
Towards the toe of the slope, the thickness flow measurements. The majority of the drains
of altered material reduces. Underlying the are drilled in the middle sections of the slope.
argillic altered porphyry is silicified por- Generally, they are drilled horizontally to a
phyry. These units are termed the altered length of 300 m. Drain flow is piped to sumps and
and silicified porphyry, respectively. pumped out of the pit.
• During drilling, airlift flows are not Regular monitoring of the drain programme
recorded above the altered-silicified rock shows that there is a relatively rapid decline in
interface. Airlift flows are variably recorded total drain flow as the drains remove water stored
from transmissive structures below the within the rock in their area of influence. High
altered-silicified rock interface. However, combined flows are only sustained where new
significant transmissive structures are not drains are continually being drilled into zones that
evident in any of the four packer test holes. have not yet been depressurized. When drains

110 NORTH WALL DEPRESSURIZATION, ESCONDIDA COPPER MINE

Fig. 4. Packer test results.

are not replaced, then there is a steady drop off in east is occurring, the initial high flows from each
total flow. This is demonstrated in Fig. 5, which drain are typically more sustained.
shows combined flow from the north wall drains
from 1996 to 2000. The drain data infer that the Current pore pressures
amount of active recharge to the slope is minimal. A selection of hydrographs are displayed in Fig. 6.
In other sectors of the pit where recharge from the The graphs show that the east wall interceptor

Fig. 5. North wall combined drain flow.

P. McKELVEY ETAL 111

wellfield is starting to cut off recharge flow to the
slope. As PO-58 and PO-14 show, heads at the
top of the wall were static or rising but are
now dropping. PO-51 confirms the impact of
the adjacent pumping well, PW-47. In December
1999, there was a head reduction of up to 20m
when PW-47 commenced pumping. When the
well is shut off, there is a very quick recovery. In
general, over most of the slope, there is
a background pore pressure reduction rate of
about 0.5m month" . Therefore, under current
conditions, a reduction in pore pressure of about
6 m year"l would be expected.
Data from the northeast section of the wall
confirm that the interceptor system is beginning
to cut off recharge to the north wall, with an
accompanying fall in heads. Boreholes on the
central and western sections of the wall show that
there is a steady background reduction in pore
pressure caused by the existing drains and by
passive seepage. Higher drawdowns are created
in the hydraulic compartments around pumping
wells.
A series of pore pressure profiles were drawn
over the wall, locations are as displayed in Fig. 3.
The profiles were drawn to provide pore pressure
data for the geotechnical analysis of the north
wall. Three of the sections are displayed in Fig. 7
and described below. For consistency, heads in
the altered material are described as the 'phreatic
surface'; the heads in the unaltered or silicified
rock are described as the 'piezometric surface'.

P4
This section displays the hydrogeology from just
to the east of the Panadero Fault. The section
shows that to the west of the Panadero Zone,
altered Escondida Porphyry is underlain by
silicified porphyry. The piezometric surface in
the unaltered material at the top of the slope is
approximately SOmbgl. The phreatic surface in
the altered material is 20m lower than this.
Therefore, there is an upward head gradient from
silicified rock to altered porphyry. This trend
continues towards the bottom of the slope. At
the toe of the slope, the head in the silicified rock
is at ground level. The altered material is very thin
or absent. Significant flows are encountered in
silicified material at shallower depths. The flows
may be related to fracture zones that have been
enhanced by stress relief and rock mass
unloading.

P6
The pattern in this section is broadly similar to
Fig. 6. Water level hydrographs. that seen in the previous section. The section

112 NORTH WALL DEPRESSURIZATION, ESCONDIDA COPPER MINE

Fig. 7. Pore pressure cross-sections.

shows that the north wall is separated from surface in the altered material is closer to ground
the rhyolites to the east by the hydraulic barrier level at 42mbgl. However, artesian pressures
of the Panadero system. To the west of the were encountered in the silicified rock, and the
Panadero, altered material overlies silicified piezometric surface in this unit is approximately
porphyry. The altered material is thickest 30 m above ground level. Therefore, in this sector
(250 m) at the top of the slope and thins towards of the slope there is an upward head difference of
the bottom. at least 70 m.
In the top of the slope, the piezometric level in The influence of the existing horizontal drains
the silicified rock is approximately 50 m above the in the altered material can be seen in this section.
phreatic surface in the altered material. This In the bottom section of the slope, the phreatic
pattern is also seen in the middle slope section, surface is relatively flat because of depressuriza-
and is confirmed by packer tests and multilevel tion caused by the horizontal drains. Towards the
piezometers. The altered material in this sector is upper middle section of the slope, beyond the
approximately 250 m thick. The phreatic surface zone of influence of 300 m drains, the phreatic
in this unit is at SOmbgl. However, the piezo- surface becomes steeper. At the toe of the slope,
metric surface in the silicified rock is 40 m higher the piezometric surface is at ground level or
than this. Further down the slope, the phreatic slightly above.

P. McKELVEY ETAL 113

P9 compartmentalized by faulting, both laterally
and vertically. Intersection of any given fracture
In this cross-section there is only minimal active
will only allow drainage of the discrete hydraulic
recharge in the rhyolites behind the crest of the pit.
block that contains the fracture. Therefore, to
Consequently, heads within the rhyolites north of
achieve reasonable drainage of the overall pit
the pit are low at around 160mbgl. Heads in the
slopes, fractures at multiple locations will have to
deeper rhyolite layer are approximately 25m
higher than measured in the shallow rhyolite layer. be encountered and drained.
The unfractured rock mass has very low
At the top of the slope, in the Escondida
drainable porosity (less than 0.1%). Virtually all
Porphyry, the phreatic surface in the altered
material is approximately lOOmbgl. The piezo- of the water is contained in fractures (the main
fracture zones and interconnected microfracture
metric surface in the silicified rock is 25 m higher
systems). Main fracture zones may typically yield
than this. This upwards head gradient is
maintained through the middle slope and to the 0.1 -3.01 s~ ! , sufficient to sustain flows in vertical
toe, with the pore pressure in both units tending wells and horizontal drains. Where drains in the
to become closer to ground level towards the silicified rock intersect such transmissive frac-
tures, high yields can be obtained. An example of
bottom of the pit. At the bottom of the pit, the
piezometric surface in the silicified rock is close this is the recent pilot hole for a drainage tunnel.
The yield of the drain was 21s" 1 at a distance of
to ground level.
200 m into the slope.
Water in the silicified porphyry is currently
Conceptual model of groundwater flow leaking upward into the altered porphyry. The
water is being removed from the altered material
A conceptual model was prepared to assist in by horizontal drains and seepage to the face.
numerical modelling of the slope; the components Vertical wells, constructed in the silicified
comprise the rhyolites outside the pit and the porphyry, drain storage within relatively small
Escondida Porphyry within the pit as follows hydraulic compartments. Nearby piezometers
(Water Management Consultants 2000). react quickly to the onset of pumping. Obser-
vation boreholes at greater distances do not record
Rhyolites any obvious effects from well pumping.
Pilot hole drilling on the north ramp has not
On the northern margin of the pit, rhyolite of yielded any sites with the required fractures and
intermediate permeability overlies very low airlift flows to justify the construction of
permeability rhyolite and andesite. The hydro- additional production wells. The reasons for the
geology within the rhyolite is compartmentalized lack of productive fractures may be that the wells
by the steeply dipping NW-SE faults. These only intersect approximately 40m of unaltered
create barriers to southward flow within the rock below the altered material. Therefore, not
rhyolite towards the pit, so that continuous enough of the silicified porphyry has been
major seepage within the rhyolite along the north 'sampled'. In addition, the vertical wells may
wall is not observed. not be intersecting the steeply dipping structures
Groundwater heads are significantly below at the head of the slope.
ground level, up to 160mbgl. Active recharge
from the Hamburgo Basin to this zone is partially
blocked by the Panadero flow barrier. Therefore, Altered (argillic) porphyry
depressurization is not required in the rhyolites to Current piezometric heads in the altered porphyry
the north of the north wall. are typically 20-60m lower than in the
underlying silicified porphyry. Therefore, over
most of the slope, there is a strong upward
Silicified Escondida porphyry
hydraulic gradient. Heads are typically 50-150 m
Heads are typically 100m below topographic below topographic level beneath the crest of the
level beneath the crest of the slope, up to 40m slope and are typically 50–60 m below topo-
above, at mid-slope, and at or near the graphic level at mid-slope. The altered porphyry
topographic level near the toe of the slopes. is typically thin to absent near the current toe of
The matrix permeability of the silicified the slopes, but, where present, the phreatic surface
porphyry is less than 10~ 8 cms~ 1 . However, the is less than lOmbgl.
unaltered porphyry does contain discrete open The matrix permeability of the altered por-
fractures, which will allow it to be drained and phyry is typically between 10-4 and 10-5 cm s~ ! .
depressurized. The fracture flow groundwater Although it is still controlled to some extent by
system in the unaltered porphyry is highly relict structures, groundwater flow in the altered

114 NORTH WALL DEPRESSURIZATION, ESCONDIDA COPPER MINE

porphyry is much more homogenous than flow in each cell (tangential to the radial direction)
the unaltered porphyry. The altered porphyry does depends on the distance of that cell from
not contain any discrete high-permeability the origin of the axi-symmetric model.
fracture zones. These have become 'healed' as • As MODFLOW is designed for modelling
part of the alteration and weathering process. The flow in systems with Euclidean geometry
unit behaves more like a 'sponge'. Compartmen- (using an x-y-z co-ordinate system), the
talization is much less important than in the model parameters had to be calculated in
unaltered Escondida Porphyry. such a way that would enable MOD-
The altered porphyry has a drainable porosity FLOW to represent a system with
typically within the range 0.1 -0.5%. Water in the axi-symmetric geometry (an r-9-z co-
upper layers of altered porphyry is draining ordinate system).
mainly to horizontal drains. Drains can yield an
average of 0.11 s~' in the altered material. Com- Starting from initial conditions with ground-
bined flows amount to 2-3 Is" 1 when there is water heads set uniformly to 3000 masl, the
active drain construction. However, flows of up to model simulates 2 years during which water is
11s"1, necessary for an effective vertical well, are removed solely by evaporation from the pit wall
unlikely to be sustained. In addition, there is slow surface. This is followed by a further 2 years,
lateral seepage to the pit face. during which additional water is removed via the
horizontal drains. At this point, the modelled
groundwater heads were compared to the
Numerical modelling observed head profile along the cross-sections.
The match between the modelled and observed
The purpose of the model was to allow the head profiles on section P6 is shown in Fig. 8.
simulation of a number of depressurization This figure demonstrates that there is a good fit;
options and, therefore, to aid in the design of the in particular the upward head difference
most effective depressurization system for the throughout the slope is replicated.
north wall. The construction of the model is The water balance for the calibration run
summarized as follows: shows that:
• The flow system was modelled using the
MODFLOW finite-difference code (McDo- • Recharge to the slope from the constant
nald & Harbaugh 1988), via the Ground- heads behind the crest is of the order of
water Vistas pre/post-processing package, 0.11 s ~ l . This replicates the situation on the
and a specially constructed spreadsheet. north wall west of the Panadero where there
• A two-dimensional, axi-symmetric, verti- is very little recharge from the rhyolite to
cal slice model was constructed based on the north.
hydrogeological section P6. The model • The initial high evaporation (EVT)flowsof
simulates the flow conditions in a sector of 101s"1 reduce to 21s" 1 as the phreatic
the northwest wall of the pit, centred on surface is drawn back from the face and
section P6 (Fig. 3). evaporation is reduced. There is a further
• In modelling the flow system as axi- reduction in EVT when the drains are constr-
symmetric, it is assumed that the dominant ucted and the phreatic surface drops further.
flow directions are radial flow towards the • When the drains are constructed, there is an
centre of the pit, and vertical flow between initial high drain flow of up to 101s" 1 .
the silicified and altered Escondida Por- After 2 years of operation, combined drain
phyry. Flow parallel to the pit slope is flow is of the order of 2.51 s ~ ] . This com-
assumed to be negligible. pares well to the current combined drain
• The fracture network in the porphyry, flow of 2-3 Is" 1 .
through which groundwater flows, has been • Average flow from the silicified porphyry
represented by an 'equivalent porous medi- to the overlying altered porphyry is of the
um'. This is a simplification of the field order of 1-2 Is" 1 . This flow is driven by
situation, which is based on the assumption the upward head difference between the
that the occurrence of water-bearing two units. The current drains are con-
fractures does not vary greatly from one structed mainly in the altered porphyry so
model cell to another. the phreatic surface drops faster than the
• The groundwater model consists of a piezometric surface. Therefore, it can be
single layer with 48 rows and 96 columns. inferred that the upward head difference
Each model cell represents a volume has reduced over the last few years and the
12.5m high by 12.5m long. The width of upward flow has diminished.

P. McKELVEY ETAL 115

Fig. 8. Model calibration.

• It can be concluded that any depressuriza- If the existing drains had not been
tion system needs to stop 1-2Is" 1 of installed, the phreatic surface would have
upward recharge flow from the unaltered been close to the face of the slope with
to altered porphyry, as well as removing discharge via evaporation and the piezo-
stored water. metric surface would have been up to 30 m
above ground level.
The numerical model was used to assess a Currently the upward head difference
number of depressurization options. At this stage, drives a flow of approximately 1-2Is" 1
each option has been simulated for a two-year from the silicified rock into the altered
period. The model results represent conditions at material.
the end of two years. Except for the first case, the Wells at the base of the pit initially
prediction runs assume that the current horizontal produce significant local drawdown at the
drain network remains in operation, in addition to toe of the slope but no significant change
the simulated new drainage measures. The depres- to the pressures in the top and middle
surization options that were modelled in the sectors of the slope.
predictive runs include: Wells at the top of the slope have a local
effect on the piezometric level but the
• conditions without the existing north wall target is only achieved up to a distance
horizontal drains in the altered material; of approximately 100m either side of the
• wells in the silicified rock at the base of the wells.
pit; When drains are drilled into the silicified
• wells in the silicified rock towards the top rock there is significant additional
of the slope; reduction of the piezometric and phreatic
• drains in the silicified rock; surfaces in the middle sector of the slope.
• a drainage tunnel in the silicified rock with The piezometric surface target is achieved
various combinations and spacings of over the area of coverage of the drains, i.e.
horizontal, subvertical and vertical drains. 500m into the slope.
The cross-sections in Figs 9 and 10 present a A drainage tunnel with horizontal and
selection of the modelled piezometric and subvertical drains produces drawdowns
phreatic surfaces. These are presented in relation at the top and middle of the slope, of
to the target pressures for the slope, produced by 50-100m, which are significantly higher
a geotechnical analysis. The modelling results than all other cases. Upward flow from the
show the following: unaltered to altered porphyry is stopped and

116 NORTH WALL DEPRESSURIZATION, ESCONDIDA COPPER MINE

Fig. 9. (a) Modelled heads with no active pressurization. (b) Modelled heads with vertical wells.

the targets are achieved at distances of up to tal drains in this unit will become more inefficient
800 m from the toe of the slope, i.e. over the as the phreatic surface drops, they are still the
length of coverage of the tunnel system. most effective method of reducing the phreatic
surface. Drain drilling in the altered material
The depressurization system that has the most needs to be implemented on a continual basis, so
significant effect on all sectors of the slope is the that 2-31s"1 is discharged from that section of
tunnel option with horizontal and sub vertical the slope.
drains. The model indicates that this layout pro-
duces pressure reductions that are up to 15 times
those currently measured. The targets for the Depressurization system design
slopes are effectively achieved with this option. The analysis and modelling demonstrates that an
An examination of the water balance from the effective depressurization system for the north
tunnel prediction shows that the tunnel stops and NE walls should consist of a combination of
the upward flow of 1-2Is" 1 from the silicified elements, as follows.
rock to the altered material. However, the pore
pressure reduction in the silicified porphyry is not
sufficient to induce flow downwards from the Elements
altered material. Therefore, active drainage Recharge interception. The modelling assumes
measures will still be required to remove stored that there is no significant recharge flow from the
water in the altered porphyry. Although horizon- Hamburgo Basin. The constant heads on the left-

to the north of the pit. Horizontal from the slope. This is because the current drain important that the interceptor wellfield along the network has substantially dewatered the altered east wall of the pit will continue to be operated as material. 2–31s" 1 is . be continued so that. tive method of reducing the phreatic level. therefore. hand boundary of the model contribute approxi. Drainage tunnel. (a) Modelled heads with drains in silicified porphyry. Drain drilling should effects on heads at the toe of the slope. as currently constructed in the altered flow from the east. P. This sive lateral drain network will produce the represents the conditions west of the Panadero. but will not cause downward the base of the pit. will be a part of the future depressuriza- pressure targets will be achieved. tion system. then it is unlikely that pore material. on average. and horizontal drains are the most effec- planned. required depressurization of the silicified where there is minimal recharge in the rhyolites porphyry. A 500 m tunnel with an exten- mately O . Active drainage surface at the bottom of the pit below the will still be required in the altered porphyry working pit floor. The tunnel will reduce the piezometric surface in the silicified rock and stop upward leakage to the Vertical wells. 10. These will reduce the phreatic flow from the altered material. The planned depressurization methods will remove a long-term maximum of 10-15 Is" 1 Horizontal drains from the face. (b) Modelled heads with drainage tunnel. Vertical wells will be required at altered material. l l s .1 to the model domain. If there is significant recharge drains. McKELVEY ETAL 117 Fig. It is. and have significant local to remove stored water.

There. used where appropriate. would like to thank the staff of MEL and WMC who meters. plus 50m. we and completion techniques for drains and piezo.118 NORTH WALL DEPRESSURIZATION. Tunnel location. vertical wells and drains from the face. above elements.8 X 3. By removing small amounts of water over fore. the tunnel and associated drains will meters. the optimal procedures for the sectors of the slope. the phreatic surface is close to ground level and The total meterage of tunnel drains has been where new drains will have a significant impact. 45° from horizontal) drilled from the last 200 m of the tunnel to drain the increased thickness of Tunnel design saturated material above the tunnel. with stations for drain drilling every The groundwater model of the north wall. optimized by using the model. 11. It is estimated The depressurization programme for the north that a maximum of 7000-10 000 m of drains will wall will be an interactive combination of the be required. such as targeted A preliminary test phase is also included. There are been evaluated. tunnel.2 m. The tunnel location is shown in Fig. The implementation of each In general. discharged from the altered material. it will be driven upwards at an angle of 2° from a portal Conclusions near the base of the slope. . Its dimensions will be 3. a large area. and construction for permission to publish this work. as the tunnel is being constructed. for example near PO-93. and (2) subvertical drains (up to review and re-evaluation. has shown that the most effective means of reducing pore pressures in the An intensive monitoring programme will be deep silicified Escondida Porphyry is a drainage carried out to allow interactive decisions to be made on each subsequent drilling phase. 11. there will be two types of drain: element will need to be flexible and based on the (1) horizontal drains drilled laterally from all results of the monitoring. and on continual drilling stations. The test will investigate the effect of varying the We are grateful to Minera Escondida Limitada (MEL) length and orientation of drains. a number of piezometers will be bring pore pressures below the specified targets constructed from the tunnel to measure pore and will optimize the stability of the wall. ESCONDIDA COPPER MINE Fig. in addition to the existing vertical piezo. The pressures in areas which are inaccessible from depressurization of the slope will be an interactive the surface. using drains drilled from the first two drilling stations. Once the preliminary testing phase has helped in the study. with each drain up to 300m long. process with additional elements. practical field experience. where drain programme will be determined. In addition.

Consultants. Modular Three Dimensional Finite-difference Water Management Consultants. M. Shrewsbury. Escondida References Mine Interim Dewatering Status Report.G & HARBAUGH. Shrewsbury. A. 1999. Water MCDONALD. 1988. UK. McKELVEY ET AL 119 Water Management Consultants. 2000. . US Geological Survey. P. UK. Escondida Ground-water Flow Model. Water Management 06-A1. Mine.W. North Wall Hydrogeology and Design of a Techniques of Water Resource Investigations. A Management Consultants. Depressurisation System.

Hydrogeological framework for assessing the possible environmental
impacts of large-scale gold mines
J. S. KUMA11,2 P. L. YOUNGER2 & R. J. BOWELL3
1
KNUST School of Mines, PO Box 237, Tarkwa, Ghana (e-mail: j.s.y.kuma@nclac.uk)
2
Department of Civil Engineering, University of Newcastle, Newcastle upon Tyne NE1 7RU, UK
3
SRK, Summit House, 9-10 Windsor Place, Cardiff CF10 3RS, UK

Abstract: Hydrogeological information is crucial to the development of a sound environ-
mental impact assessment (EIA) for a proposed mine, as well as the management of
potential environmental impacts during and after exploitation. However, the determination
of hydrogeological parameters is not customarily included in mineral exploration surveys,
with the result that many EIAs end up being rather light in hydrogeological content.
Examples from the Tarkwa gold mining district of Ghana illustrate this point. Consequences
of such an inadequate hydrogeological understanding are potentially serious, ranging from
an inability to predict future problems in water quality after the cessation of mining, to a lack
of understanding of hydrogeological controls on slope stability, which is arguably manifest
in the catastrophic spill of cyanide-rich processing effluents from a breached tailings dam at
Wassa West, near Tarkwa, on 16 October 2001. To redress this deficiency, we propose that a
hydrogeological database be assembled during the mineral exploration phase, according to a
specified protocol ('check-list'). Using these data, a rational conceptual hydrogeological
model for the mine site and its surrounding area can be developed, providing the basis for a
thorough consideration of groundwater aspects within the statutory Environmental Impact
Assessment, which is (as in most other countries) required by Ghanaian government statute
before a mining lease is approved. The resources required to set-up such a database are small
compared to the benefits.

Mineral exploration and exploitation is a impact assessment (EIA) (e.g. Pettyjohn 1985;
multimillion dollar business. Even though the Kolm 1996; Stone 1999).
risks are very high, the quest for improved living This paper illustrates the hydrogeological
standards and developments in mineral bene- shortcomings of EIA procedures using examples
ficiation technology both stimulate the explora- from the Tarwka gold mining district of Ghana.
tion and development of new sources of mineral A protocol for hydrogeological data collection
wealth (Woodall 1984). The mining industry to improve the insertion of groundwater system
is both an important source of employment and concepts into future site appraisal and manage-
a major foreign exchange earner in many coun- ment activities is proposed.
tries, and this is certainly the case in Ghana, the
former 'Gold Coast' of Africa. Although
environmental controls during active mining
are increasingly stringent, former mining Gold mining and EIA requirements
operations have left a long-term legacy of en- in Ghana
vironmental problems on almost all continents
(e.g. Hedin et al 1994; Younger 2001; Ghanaian gold mining and EIA requirements
Dzigbodi-Adzimah 1996). Many of these long- In Ghana, gold mining contributes approximately
term problems arise because of a lack of 30% of the annual foreign exchange earnings and
appreciation of the hydrogeological setting of more than 2.3 million ounces of gold is produced
the mine, resulting in the pursuit of inadequate by the country annually. Figure 1 shows a sim-
closure strategies. This lack of hydrogeological plified geological map of southwest Ghana and
appreciation belies the existence of well- the location of large-scale gold mines. A new
established techniques for relatively low-cost mining permit can only be issued after the En-
development of conceptual hydrogeological vironmental Protection Agency (EPA) of Ghana
models to underpin the process of environmental has approved an EIA report. The EIA report

From: YOUNGER, P.L. & ROBINS, N.S. (eds) 2002. Mine Water Hydrogeology and Geochemistry. 121
Geological Society, London, Special Publications, 198, 121-136. 0305-8719/027$ 15.00
© The Geological Society of London 2002.

122 J. S. KUMA ETAL.

Fig. 1. Simplified geological map of Ghana showing the location of large-scale gold mines (adapted from
Eisenlohr 1992).

includes, among other requirements, an Environ- Examples of hydro geological shortcomings
mental Baseline Study, Environmental Impact in Ghanaian mining EIAs
Statement and an Environmental Management
Plan. On-going mining activities are required to Even though the EPA requires water sections
submit Environmental Action Plans (EAP) with in EIA studies, many such studies lack sufficient
an annual report (Anon. 1994). During mine hydrogeological detail to permit full scrutiny
operation the EPA also administers a random of the intentions of the mining lease permit appli-
monitoring programme. cants. Without such detail, the hydrodynamics

HYDROLOGICAL FRAMEWORK FOR GOLD MINE ASSESSMENT 123

of the mining areas cannot be conceptualized to network in the Tarkwa area that would be needed
the degree necessary to allow adequate safe- to independently assess the likely rates of filling
guards to be put in place to forestall damage of of post-mining pit lakes, and/or the positions and
water resources. There are even times when little quality of mine water decants to the surface
or no hydrogeological information is presented. environment. Without such information, preven-
For example, one recent EIA reported that 'very tative measures cannot be devised (cf. Younger &
little hydrogeological information was available. Adams 1999; Shevenell 2000). Even before mine
Groundwater seems to be abundant as there was closure, groundwater data are critically import-
the need for dewatering before blasting could be ant in the design of safe pit wall angles, and
conducted by the mine' (Amegbey 1996). Even secure bunds for tailings dams and other stru-
in situations where some hydrogeological ctures (see Younger et al 2002). With regard to
information is presented, important aspects are the latter, inadequate control of sub-dam pore
normally absent from the document. For instance water pressures is suspected to have played a role
an internal report intimated that 'no ground water in the partial failure of a tailings dam, which gave
recharge information for the area exists' (Anon. rise to a release of cyanide-rich water in June
1997). 1997 (spill from the Teberebie gold mine into the
The ultimate consequence of a lack of hydro- Awunabeng stream) and perhaps also contributed
geological data is an inability to manage en- to a similar spill (from Goldfields of Ghana
vironmental and geotechnical processes that gold mine) on 16 October 2001 that introduced
influence/are influenced by groundwater occur- cyanide into the River Essuman, which provides
rence, movement and quality. Hence, there drinking water to the villages of Abekroase and
still exists no regional groundwater monitoring Huniso (Fig. 2).

Fig. 2. Geological map of Tarkwa area (adapted from Junner et al. 1942).

124 J. S. KUMA ETAL

Hence, both short- and long-term environ- be closely considered during hydrogeochemical
mental management requries hydrogeological studies (Edmunds 1981). During these studies
understanding. Such an understanding does not the type of clay present should be determined by
necessarily require extensive groundwater map- X-ray diffraction (XRD) if possible, as this
ping to 'First World' standards and it can, in fact, information can improve knowledge of the
be achieved using an array of relatively simple possible chemical reactions likely to occur in
methods. the shallow subsurface (Appelo & Postma 1993;
Head 1997). In addition, the rainfall character-
istics, vegetation and land use all influence in-
Protocol for low-cost hydrogeological data filtration of water through soil, and form a critical
collection for inclusion in EIA component of the hydrological cycle (Dingman
1994; Dunne & Leopold 1998).
Groundwater information can be gathered during In undertaking a pedological survey the soil
the phase of mineral exploration, so that it is thickness encountered during drilling should be
readily available at the time the EIA is under- recorded, to delineate the weathered-unweath-
taken. Although the following items are des- ered boundary of the area. Furthermore, the soil
cribed in a segmented manner, it should be borne profile and any peculiar textural and structural
in mind that they are in reality interconnected. characteristics in all excavations should be
Hence, some of the items covered in one section logged. Soil should be sampled in the B-horizon
below are also highlighted in subsequent because this is the most critical horizon on which
sections. infiltration of precipitation to the groundwater
zone depends (Davis & DeWiest 1991). An
Physiography optimum of 40 soil samples is recommended for
determination of particle size distribution (PSD)
Historical information on the physiography of analysis, bulk and particle densities, and moi-
the area may be available, but data on current sture content tests. Soil texture, sorting, changes
land use, vegetation, relief and drainage infor- in lithology, volume changes as a result of
mation are normally collected early in the pre- compaction and subsidence, and porosity are de-
liminary exploration phase by conventional termined from these tests. Moreover parameters
surveying. A map of the physiography at a such as porosity, permeability and heterogeneity,
scale of 1:20 000 or less is normally prepared for together with shape and size of voids, are likely
the concession, and this scale is reduced as to be of considerable importance in controlling
detailed phases of exploration are pursued. the movement of pollutants through the unsatu-
Surface water systems provide valuable evidence rated zone (Hounslow 1985). Assessment of
of groundwater movement (where the two groundwater recharge is also aided by using this
systems are hydraulically connected) and ana- information.
lyses of the surface water system will enable
information to be obtained about groundwater
flow patterns and water quality (Fetter 1994; Geology. Knowledge of the lithology, strati-
Winter et al. 1998). In addition, the drainage graphy and structure of the rocks in a region is
pattern of an area is normally a reflection of its essential to understanding the nature and dis-
geological history and structure (Stone 1999). tribution of their water-bearing properties (Fetter
Thus, due consideration given to the study of the 1994). In a mining environment the presence of
drainage pattern will improve an early con- lithological logs will greatly enhance knowledge
ception of the groundwater system in the area, of the subsurface geology, which in turn helps to
for example, identification of recharge and dis- improve conceptions of the hydrogeology of the
charge areas. area.
Lithological logs are collected during the
exploration phase. Normally during the stage of
Pedology detailed exploration, these logs are acquired at
The physical and chemical characteristics of a spacings as small as 25 m and in a regular
soil play a vital role in the recharge and chemical pattern, at least in and around the ore body. These
evolution of groundwater. This is because soil logs should be scrutinized for all the hydro-
texture and structure control hydraulic proper- geologically relevant information that they may
ties, whereas soil mineralogy affects water contain by means of logging and laboratory tests,
quality (Hem 1992; Stone 1999). It has, there- which can then form the basis of a hydro-
fore, been advocated that the nature of the geological model of the concession. Younger
soil and the drainage characteristics of an area (1992) has demonstrated the use of a petrological

HYDROLOGICAL FRAMEWORK FOR GOLD MINE ASSESSMENT 125

microscopy in the direct measurement of grain can also be estimated and compared with
size, pore size and porosity, from which estimates the value from the stream hydrographs. These pie-
of hydraulic conductivity, specific yield and zometers must be kept in working order
solute retardation factors have been made. It has throughout the life of the mine so that future
also been shown that the storativity of an aquifer assessments can be conducted without difficulty.
can be estimated as function of aquifer lithology
and thickness alone (Younger 1993). A probe
permeameter could also be employed for making Hydro geochemistry
inexpensive permeability determinations on Surveys of streams and springs in and around
borehole core samples (Eijpe & Weber 1971). the concession area, particularly during dry
Geological logging and petrological studies of periods when there is no surface run-off to mask
cores enable the prediction of water–rock inter- groundwater contributions, can provide invalu-
actions likely to explain the chemistry of ground- able information on groundwater chemistry (e.g.
water discharges to streams (Sharp & McBride Pettyjohn 1985). The data also provide a useful
1989; Appelo & Postma 1993). The presence of input to the baseline information required in EIA
sulphide minerals, which are prevalent in some reports. However, where such surveys are under-
gold ore fields, deserves particular attention due to taken, they often include analysis of only the
their propensity to release acidic, metalliferous major ions and/or samples prove to have been
leachates (acid mine drainage) after they have poorly collected, so that incomplete and/or
been aerated. Early knowledge of their occurrence contradictory information is provided. We pro-
is useful during planning. pose the following list of determinands: tem-
There is a paucity of information regarding perature, pH, Eh, Na, Ca, Mg, K, HCO3, SO4, Cl,
the aquifer properties of many mine concessions, N03, Si02, Fe, Mn, Al, Co, Cr, Pb, Zn, Cu, CN,
even though water boreholes are drilled for Hg, As, DO, TDS colour and turbidity. Further, a
processing and other needs (e.g. for local major omission is often the assessment of pit-
communities on mine concessions and for those wall or waste rock contribution to discharge
who have been resettled) during the feasibility water quality and changes that will occur over
stage of exploration. A little time and effort time (Bowell 2002). This information is nor-
invested in pumping tests of water boreholes mally not collected because an experienced
drilled for such purpose can be an invaluable hydrogeochemist may not be included in the
addition to the hydrogeological database. Some exploration team.
of the exploration boreholes can also be adapted At least 5 years (some times up to 20 years)
as monitoring boreholes to establish pre-mining elapse before an exploration prospect actually
groundwater trends and background hydroche- becomes a mine (Woodall 1984). Therefore, if all
mical patterns. the aspects above are conscientiously followed
during the exploration stage, a sufficient run of
Hydrology data will have been accumulated to support
development of a robust conceptual hydrogeolo-
Information on the climate may be available gical model of the mine.
for the region as a whole. However part of
the concession may have been excluded from
further consideration (typically 50% or more). Example application of the protocol
Rainfall, evaporation, temperature and other to the Tarkwa gold mining district
meteorological data are normally a requirement
for site design. Therefore, a simple meteorologi- To test the feasibility of applying the above pro-
cal station is often installed to aid the estimation tocol to a real system, it was applied to the
of evapotranspiration. Evaporation measurements Tarkwa mining district in Ghana by means of
are also necessary for the design and operation field surveys between January 2000 and January
of heap-leach pads, which are employed by most 2001. The following sections document the
of the gold mines. It is advisable to start routine findings.
gauging of major streams in the area (for run-off
and recharge estimates) at this time so that this
information can be synchronized with meteor- Physiography
ological data for future analyses and interpret- The Tarkwa area has a humid tropical climate
ation. In addition, piezometers need to be with an average annual rainfall of over 1750 mm.
economically installed at this time at well-chosen Rain falls in two main periods in the year: a major
sites in the concession so that the flow regime wet season from April to July (with a peak in
of the ground water can be determined. Recharge June) and a second minor season from September

126 J. S. KUMAETAL

to November. Air temperature for the area varies Younger 2001). The soils are mainly silty-sands,
in a narrow range (between 28 and 33°C) and which exhibit Ks values in the 10 - 5 -10 - 8 ms" 1
relative humidity varies from 83 to 91 %. Located range, although this is dominated by those in the
in an area transitional between the rain forest narrower band of 10~ 6 -10~ 7 ms~ 1 . Minor later-
and moist semi-deciduous forest, zones of the itic patches are located on hilly terrain underlain
original vegetation are locally present, while by Banket Series and Tarkwa Phyllite rocks.
other areas have been cleared for farms, mines Banket soils exhibit the most favourable
and communities (Dickson & Benneh 1988). characteristics for infiltration in the area, in
The Tarkwa district is located within the terms of both PSD and KS values. These soils are
Ankobra River basin and is bordered to the west located on hills and, therefore, act as the main
by the southerly flowing Ankobra River (Fig. 2). areas for groundwater recharge. Huni and
Both the Huni and Bonsa rivers are major Kawere soils, located in low-lying areas, display
tributaries to the Ankobra, and border the area to characteristics suggesting that the Huni, with a
the north and south, respectively. The area is much better sorting coefficient and Ks value,
highly dissected and of moderate relief with a will admit more precipitation for recharge
general decrease in hilltop altitudes towards compared to the Kawere. Soil pH varies from
the south. A series of parallel ridges and valleys very acidic to moderately acidic, that is
oriented along the general NE-SW strike of 1.72-5.01. The lowest pH values are associated
the rocks define the landscape. This geomor- with weathered felsic dykes and are caused by
phology is due to pitching fold structures and the presence of well-disseminated fine pyrite
dip-and-scarp slopes of the Banket Series and crystals, which are expected to be a potent
Tarkwa Phyllites, which form part of the bedrock source for acid rock drainage when found in the
(see below). Transverse to the ridges and valleys unsaturated zone.
are smaller tributary valleys and gaps controlled
by faulting and jointing (Whitelaw 1929).
Geology
Pedological characterization Sediments of the Tarkwaian System are predo-
Infiltration and particle size distribution (PSD) minantly arenaceous and were 'deposited by
tests were conducted to determine saturated high-energy alluvial fans entering a steep sided
hydraulic conductivities (Ks) and textural charac- basin filled with fresh water' (Kesse 1985). They
teristics, respectively, of soil at 56 sites in the consist, in general, of coarse, poorly sorted,
Tarkwa area (Kuma & Younger 2001). These soil immature sediments with low roundness typical
tests were conducted in the B-horizon (Table 1). It of a braided stream environment. They have
was observed, in general, that extremely poorly been metamorphosed to low-grade green-schist
sorted soils exhibit low porosity and relatively facies (Kesse 1985). Rocks of the Tarkwaian
higher KS while relatively better-sorted soils System, in the direction of younging, comprise
reveal high porosity and low Ks values (Kuma & the Kawere Group, the Banket Series, the Tarkwa

Table 1. Summary of soil tests conducted in the Tarkwa area

Percentage Ks(10-6ms-1)
Soil type Texture
Gravel Sand Silt Clay Min. Max. Mean 50 W% Gs pH
Banket Silty-sand 2 59 29 10 0.21 10.56 4.45 3.74 0.38 17.5 2.66 4.92
Laterite 69 14 10 7 0.01 20.75 3.46 5.85 0.22 9.8 2.67 4.86
Huni Silty-sand 2 55 33 10 0.05 7.28 1.57 2.28 0.42 12.8 2.65 5.01
Kawere Silt sand 0 47 40 13 0.10 1.87 0.72 4.34 0.38 12.4 2.65 4.65
Tarkwa Laterite 62 9 13 16 _ _ _ 17.8 - 14.2 2.74 5.01
Phyllite
W. Dyke Silt 3 20 64 13 0.31 2.57 0.88 2.75 0.40 22.2 2.62 5.22*
1.96*

Ks is saturated hydraulic conductivity, S0 sorting; n is porosity; W% is moisture content, Gs is specific gravity;
W. Dyke is weathered dyke.
*Mafic.
T
Felsic.

HYDROLOGICAL FRAMEWORK FOR GOLD MINE ASSESSMENT 127

Table 2. Subdivision of the Tarkwaian System in the Tarkwa area (modified after Junner et al. 1942). The boldface
characters under 'composite lithology' signify the most abundant and important rocks forming each division

System Series Thickness (m) Composite lithology
Tarkwaian System Huni Sandstone 1370 Sandstones, grits and quartzites with
bands of phyllite
Tarkwa Phyllite 120-400 Huni sandstone transitional beds and
greenish-grey phyllites and schists
Banket Series 120-600 Tarkwa Phyllite transitional beds and
sandstones, quartzites, grits breccias and
conglomerates
Kawere Group 250-700 Quartzites, grits, phyllites and conglomerates

Phyllite and the Huni Sandstone (Fig. 2 and where: P is the precipitation (mm year ), AE is
Table 2). the actual evapotranspiration (mm year-1), RO is
Intrusive igneous rocks make up about 20% of the catchment surface run-off (mm year- ! ), SMS
the Tarkwaian System in the Tarkwa area. These is the soil moisture storage (mm year -1 ) and SWS
rocks range from hypabyssal felsic to basic is the surface water storage (shallow lakes, mm
igneous rocks, principally in the form of conform- year- 1 ).
able to slightly transgressive sills, with a small The Penman-Pan evaporation relationship is
number of dykes. Faults and joints are common in strongest in the more humid areas of Ghana and in
the area and the most prominent joint sets trend the wet seasons, because the potential evapotran-
ESE-WNW (although NW-SE and N-S trends spiration and pan evaporation values are similar to
are also present). The faults are either strike- those yielded by the Thornthwaite and Papadalus
parallel and closely associated with folding (and formulae (as described by Acheampong 1986).
may occur as upthrusts) or dip-parallel, and are Dunne & Leopold (1998, p. 136) came to a similar
most often recognized as breaks in the topography conclusion working in Kenya (which is in the
(Hirdes & Nunoo 1994). tropics, on about the same latitude as Ghana,
With shallow-water beds in a braided environ- although higher in altitude). Considering the
ment, the thickness of individual members varies results of these studies, pan evaporation data were
considerably. In addition, a fractured and meta- adopted as suitable for describing the potential
morphosed lithology confers aquifers with dual evapotranspiration (PE) of the study area.
porosity, limited areal extent and storage pro- A method adopted by the British Meteorolo-
perties. gical Office (BMO) for estimating actual
evapotranspiration (AE) of a water basin is to
Hydrology classify the various types of vegetation according
to their root constants (RC) and proportionately
Direct recharge (sensu Lerner 1990) was ident- determine the AE from these after calculating the
ified as the dominant recharge process requiring water balance (Shaw 1994). Based on the BMO
estimation because precipitation (1803mm for model, vegetation in the Tarkwa area was classi-
the year 2000) is the primary hydrological input. fied as having about 65% mature forest (of which
In the development of the water balance model, it 25% was riparian and always transpires at the
was assumed that the area is almost hydrologi- potential rate), 10% shrubs, 15% crops, and 10%
cally closed with respect to both surface and urban and mine area. The mature forest is exp-
groundwater. This is because on three sides of the ected to have the maximum possible root constant
area, three rivers, i.e. the Bonsa, Huni and of 250 mm, based on both the soil texture and its
Ankobra, effectively act as surface and ground- vegetation (Grindley 1970). Crops farmed are
water divides with no possible intrusion from roots and cereals with approximate RC values of
beyond the river boundaries. On the eastern 100 and 150mm, respectively; shrub RC is about
boundary, the Banket ridge acts as a water divide. 200mm, and in urban and mining areas it is
The equation for precipitation recharge for the 25mm.
study area is written as: The Thornthwaite & Mather (1957) account-
ing procedure for the water balance method was
Recharge (R) = P – RO – AE SMS adopted, and Table 3 shows an example
calculation using a root constant of 250mm.
– SWS (1) The average actual evapotranspiration obtained

70 17. Pettyjohn & Henning 1979).1 . and it is highly sensitive to Three of five major streams and one smaller changes in precipitation and evaporation.33 106.97 79. Table 3. Some of these are the result of recent distortions off for the whole study area for the year 2000 was in the landscape due to surface mining while estimated at 449mm.78 -94. of mining or else have become unintentionally Fig.16.90 95.28 0 250 90 0 97 November 160.00 415.00 –16 234 84 0 0 Total 1803.80 -318. were gauged in order changes in soil moisture content occur through- to estimate their surface run-off and base flows out the year (Chidley 1981). groundwater discharge) and the surface run-off Shallow basins under water or marshy con- components (associated with precipitation) (see ditions during the dry season are significant. PL AST ST AE Deficit Surplus Month (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) January 60.80 .80 65.46 73.20 .38 11 250 74 0 6 September 123. It is generally (Fig.18 mm year.60 50. PL is the accumulated potential water loss and is defined as the cumulation of negative values of (R-PET). all in the study area.10 0 250 96 0 65 December 68. This program separates the differences may occur from one year to the next hydrograph into base flow (associated with (Shaw 1994).01 65. Those due to surface mining graph of monthly surface run-off and base flow are either abandoned pits created in the process for the four gauged catchments.00 –-42 129 79 27 0 April 166. S. 3).10.00 -11 239 86 0 0 August 91.00 0 1048 42 755 *Acc.22 84. R-PET Acc. even though significant stream hydrographs.00 –16 218 77 2 0 February 59.43 1089.67 –163.20 50.88 -11. Total surface run. KUMA ETAL. The amount by which PE and AE differ each month is the deficit (SMD). .92 102.00 –47 171 106 13 0 March 37.01 0 250 65 0 415 July 74.23 -34.30 191. Plot of discharge of four streams in the Tarkwa area. so that stream.90 -59.62 71 250 102 0 121 June 480.00 -68.92 85.90 97.33 115. capillary forces. Figure 4a-d shows the others are natural. The computer program HYSEP (Sloto & accepted that the net change in water storage on Grouse 1996) was employed to separate the an annual basis is zero.73 50 179 116 0 0 May 293.18 . AST is the change in soil moisture. Surplus is the moisture surplus.26 0 250 73 0 51 October 187.08 73.12 118.40 –16. 3.18.18 89. for the Tarkwa area using these procedures is Soil moisture (SM) is held in the soil pores by 1035. Monthly water balance at the Tarkwa area with a root constant of 250 mm * Rainfall Pan Evap.128 J.

HYDROLOGICAL FRAMEWORK FOR GOLD MINE ASSESSMENT 129

Fig. 4. (a) Graph of base flow and surface run-off of the Bediabewu catchment, (b) Graph of base flow and surface
run-off of the Essuman catchment, (c) Graph of base flow and surface run-off of the Anyuiafutu catchment, (d)
Graph of base flow and surface run-off of the Awunabeng catchment.

impounded by earth moved and tipped during stream hydrographs (using HYSEP; Sloto &
mining. Water stored in this way would pre- Grouse 1996) for the area amounted to 147mm
viously have formed part of local stream flow. year"1, i.e. about 8% of annual precipitation. For
The natural pools are surface manifestations of mountainous regions, Mau & Winter (1997)
shallow regolith aquifers. The volume of water observe that baseflow estimates could provide
present in the depressions depends on precipi- lower estimates of recharge by about 25%. The
tation and PE. The presence of water at the end of moderate relief of the Tarkwa area suggests that,
the dry season implies that PE is less than although baseflow is less than recharge, the
precipitation. A rough estimate of surface water difference is not as great as 25%. An important
storage (SWS) was made, i.e. about 20mm reason for the low value of recharge estimated
year- 1 . using stream hydrographs compared to the water
Precipitation recharge (RE) from the above balance method is that surface mining has
considerations is: distorted the topography such that some of the
rainfall is held in mined-out pits and other man-
(1803.43 - 1035.18 - 449 - 20) mm year"1 made impoundments, thereby reducing the
volume of stream discharge (as noted above).
= 299.25 mm year- 1 In addition, all three surface mines in the district
win gold from the Banket Series, which
i.e. about 17% of annual precipitation. incidentally has the highest relief in the area.
Stream hydrograph analysis, from which base- PSD and infiltration tests for the Banket soils
flow is separated, may also be used for the revealed they possess the best characteristics for
estimation of groundwater recharge (e.g. Mey- infiltration and provide many recharge areas.
boom 1961; Rorabough 1964; Bevans 1986; Thus, gold mining is inadvertently destroying
Rutledge 1992). The baseflow estimated from recharge areas for groundwater. It might be

130 J. S. KUMA ETAL.

argued that if, for example, recharge estimates • Springs and shallow wells give direct sam-
were determined ahead of the three surface gold ples of groundwater, and boreholes provide
mines commencing production, a better under- samples of bedrock groundwater. (The av-
standing of the groundwater system may have erage depth of the boreholes is 60m.)
been obtained. During these studies, the lack of depth
Baseflow index (BFI) is a catchment charac- sampling equipment meant that 'bulk' gro-
teristic and is defined as the percentage of run-off undwater samples were taken.
derived from groundwater (Anon. 1980). The
BFI determined for four catchments reveal that Table 4 summarizes results of the water
their storage characteristics are low: Awunabeng properties investigated. Spatial and temporal
(33%), Bediabewu (17%), Essumang (28%) and variations are observed, and are attributed to
Anyuiafutu (27%) (Fig. 4a-d). (Stream names mine water discharging into receiving streams
are also used to define the catchments in Fig. 2.) and local community use of the water resources.
Run-off varies with the depth and texture of the Some of the important conclusions arrived at are
soil, and physiography of the catchment (Dunne as follows.
& Leopold 1998). Therefore, these values have Different water types are observed for the
both lithological and topographic significance. different water regimes sampled. The hydro-
The Bediabewu flows more or less between the chemical facies identified with water samples
Banket and Tarkwa Phyllites before going over from springs, shallow wells and hand-dug wells
Kawere rocks. Both Banket and Tarkwa Phyllites are dominated by Na-Cl-HCO3-Cl. Streams
form narrow ridges in the area and run-off is which were perceived not to be directly or
swift, resulting in a very low baseflow (Fig. 4a). indirectly linked with mine areas are grouped as
The Essumang and Anyuiafutu streams, with 'pristine'. Pristine water samples fall into two
virtually equal BFI values, are both largely hydrochemical zones, namely, water with an
confined to the Huni Sandstone, which has a very intermediate chemical character, i.e. no cation-
gentle topography and, therefore, these values anion pair exceeds 50% and the fresh water
probably reflect a strong soil influence (Fig. 4b & facies. These water types are described by Na-
c). The Awunabeng catchment includes the Ca-Mg-HCO 3 -SO 4 and Ca-Na-HCO 3
Awunabeng and Ahumabru streams, and is chemical facies, respectively.
underlain by Banket and Kawere rocks. The Streams that have passed through and/or re-
Banket terrain is large, and is drained by ceived mine water display Na-SO4-HCO3 and
numerous tributaries of the Awunabeng stream. Na–Ca–HCO 3 –SO 4 hydrochemical facies and
Despite the fact that mining has destroyed part of are identified, respectively, with the Awunabeng
the recharge area of the Banket, the Awunabeng and Bediabewu streams. Their water chemistry
catchment still exhibited the highest BFI of the ranges from slightly saline to saline, with the
streams gauged (Fig. 4d), confirming the status highest Na sample of 373.6mgl- 1 from a
of the Banket as the best yielding aquifer tributary of Bediabewu flowing from Teberebie
lithology in the Tarkwaian System. Goldfields concession.
The Essuman stream exhibits a Ca–Na-HCO 3
Hydro geochemistry hydrochemical facies, while Anyuiafutu shows
equal concentrations of the major cations and
The following water features were identified for almost equal concentrations of SO4 and Cl, i.e.
investigation: Na–Ca–Mg–Cl–SO 4 . Low HCO3 water con-
• Dry weather water samples were collected centration associated with Anyuiafutu is likely
from all streams in the area at reach due to microbial respiration in the soil resulting in
intervals of about 3 km. The dry weather a low annual pH of 6.27. Considering that both
(or 'low-flow') period allowed any spatial Essuman and Anyiuafutu traverse the same
variations in in-coming groundwater lithological unit, the expectation is that both
chemistry to be identified (Pettyjohn will exhibit broadly similar hydrochemical facies.
1985). Differences in their water types are attributed to
• Temporal variability in stream chemistry both artificial and natural influences:
was investigated by means of periodic
gauging and sampling. All the four streams • most of the water in Essuman flows
were sampled over a 12 month period on a through mine territory and some tributaries
weekly basis for pH, Eh, temperature, receive mine water, this is not so with
bicarbonate alkalinity, Cl, Ca, SO4 and Anyuiafutu;
SiO2. The suite of major and some minor • Anyuiafutu, because of its small catch-
ions was analysed on a monthly basis. ment may be receiving only near-surface

HYDROLOGICAL FRAMEWORK FOR GOLD MINE ASSESSMENT 131

groundwater, explaining its low pH and
HCO3 values. However, the pH of
Essuman is higher than the average and
may be due to mine water discharge.

Another important factor considered during
this investigation is the ratio HCO3/SiO2.
A terrain is likely to have undergone silicate
weathering if the HCO3/SiO2 ratio is less
than 5, and carbonate weathering if the ratio is
greater than 10 (Hounslow 1995). Also, a
terrain that has undergone silicate weathering
exhibits a low total dissolved solids (TDS)
value (in the range 100-200mgr1), but in
carbonate weathered areas, TDS is typically
greater than 500 mgT 1 (Hounslow 1995).

(a) Dry weather water analyses of per-
ceived pristine areas show HCO3/SiO2
and TDS values of 2.85 and 110 mg 1-1',
respectively. These values imply that
the water in this area has participated in
silicate weathering. However, streams
passing through mining territory and
those that receive mine water show
HCO3/SiO2 ratios of 11.42 and TDS of
536mgl~ 1 , suggestive of carbonate
weathering. As the lithology in the
mines is not different from that outside
of them, this apparent inconsistency can
only be due to a direct influence of mine
water chemistry on receiving streams,
probably by disposal of alkali-dosed
waters and/or leakage of spent heap
leach pad liquors (de-cyanidized) and/
or tailings dam supernatants entering
the groundwater system.
(b) The results of (a) apply to streams
flowing through and receiving mine
water from Ghana Australian Goldfields
(GAG) and Teberebie Goldfields Lim-
ited concessions. The Essuman stream
receives discharges of mine water from
Goldfields Ghana Limited, but this
stream does not show any significant
major ion effect and its chemical facies
plot in the fresh water domain, i.e. Ca-
Na-HCO3. It is possible to speculate
that little or no leakage of unprocessed
mine water is released into the Essu-
man. However, minor and trace ion
chemical analysis should be performed
on this and all the other streams to
ascertain their complete water chem-
istry (local resources precluded this
during the present study).
(c) Mean pH of water in the pristine areas
is 6.48, but above 7.00 in streams

132 J. S. KUMA ETAL.

carrying and receiving mine water from the mineral exploration phase for new mines.
GAG and Teberebie Goldfields (it is 6.77 It has identified which datasets should be
for Essuman). The higher than average pH considered and how they can be obtained at
values appear to be due to the use of lime relatively low cost, thus easing the data
in processing gold ore, which seep into requirements during hydrogeological assessment
streams. for environmental impact assessment purposes
later in the development of the mine prospect. All
A mean TDS value of HOmgl" 1 for pristine these data are collected one way or the other
areas suggests that the residence time of during groundwater studies for other purposes. In
groundwater in the subsurface prior to their the case considered, i.e. during mineral explora-
emergence in streams is brief, and hence source tion, the problem is that emphasis is almost
concentrations are low. In addition, the low TDS entirely placed on "what grade and ore body
value also suggests a region dominated by local- magnitude are we looking at?'. Therefore, the
scale flow systems. No change in anionic people involved are either not aware that hydro-
composition from recharge to discharge points geological data are available or, if they are aware,
was observed (for streams in pristine areas), the value of data collection is not realized and it
which further corroborates small-scale, localized is seen as a waste of time and money.
water-rock interactions. It might also be argued that at an early stage
This short study of the major ion chemistry of when the geological resource is not yet fully a
water from the Tarkwa gold mining district has reserve, it would be a waste of financial resources
shown that mining has affected stream water. A and time to collect data that will not directly
long-term monitoring programme undertaken improve the reserve estimation. However, spend-
with an expanded analysis including trace and ing at this time on good quality data collection
other metal ions would enable the full impact of (which should mean employment of a qualified
mining to be stated. hydrogeologist) could offset:
Using hydrogeological data to assess • the amount to be spent for employing a
possible environmental impacts hydrogeologist during the preparation of
If all the issues raised are pursued during the that part of the EIA report. This also
exploration and planning phases of mine means that a more accurate report is
development, hydrogeological assessment for produced;
possible environmental impacts becomes rela- • the cost of hiring a consulting hydrogeol-
tively simple and may be conducted by: ogist later on when a hydrogeological
problem arises at the mine. This is because
• using the hydrogeological model in the the solution can be obtained by analysing
El A report as a baseline for assessment; the detailed report prepared earlier, or the
• making regular updates to the hydrogeo- problem will not arise anyway because the
logical database during the course of mine information available has been utilized,
operation and production. More monitor- preventing its occurrence;
ing boreholes may then be in operation • future problems (environmental, monetary
and can be sampled and gauged for water or both) due to ill-conceived planning and
quality and groundwater head studies, res- construction;
pectively; • some of the information is used as input to
• a low-flow water quality and discharge dewatering design and water management.
survey conducted in the mine concession
leading to hydrogeochemical interpret- It is now the practice to introduce environ-
ation to support the baseline report; mental studies into many academic programmes.
• field monitoring of identified poor dis- Therefore, as a permanent solution in the future,
charges or problem areas; and to cut costs, it is suggested that relevant
• hydrogeochemical studies for possible programmes for undergraduates and graduates in
amelioration. the field of earth sciences should incorporate
fundamentals of hydrogeology into their curri-
A summary of the above recommendations is
culum. This should familiarize geoscientists with
presented in Table 5 and as a flow chart in Fig. 5.
groundwater hydraulics and the groundwater
cycle. In addition to the above, an understanding
Discussion and conclusion of hydrogeological soil and rock sampling and
This study has outlined the importance of early logging procedures are essential during litho-
gathering of hydrogeological information during logical core logging. This information can be

Table 5. The recommended phases of mineral exploration most appropriate for hydro geological data collection and the expected outcomes

Hydrogeological aspect Mineral exploration phase Hydrogeological outcomes
Physiography Desk studies and during regional Maps at a scale of
• Relief reconnaissance. As more detailed surveys 1:20000 - 1:5000 are prepared during coventional
• Drainage are executed, larger-scale maps are surveying. Recharge and discharge areas
• Land use obtained. are determined. Preliminary groundwater flow
• Vegetation directions and groundwater boundaries are m
conceptualized.
Pedology Exploratory, outline and evaluation drilling Soil thickness map constructed. Soil
• Soil logging and sampling phases. Pitting and trenching are texture, sorting, porosity and changes
• PSD useful if weathering profile is in lithology determined. Results are
• Moisture content, bulk and particle thin. also useful during assessment of
densities groundwater recharge.
• Infiltration tests
Geology From the desk study phase From thin section analysis of
• Structural through to evaluation drilling, surface rock samples, determine: mineralogical composition,
• Lithological mapping and core logging. grain size, sorting and porosity
• Stratigraphic (by point counting). Use porosity
• Thin section and microscopy to estimate specific yield and
• Probe permeameter storativity (Younger 1993). Permeability determined from probe
permeameter. Prediction of rockwater interactions for assessment of
possible discharges to streams. Three-dimensional
(3-D) conceptual hydrogeological model from
above information coupled with surface
geological mapping and core logging.
Hydrology Some information is available at Evaporation and evapotranspiration estimation. Use
• Meteorological station installed for daily the desk study phase, but information from soil survey with
recording install a meteorological station, install rainfall, run-off and evapotranspiration data
• Daily gauging of major streams piezometers and start stream gauging to estimate recharge. Use stream
• Piezometers installed - all during evaluation drilling. hydrograph to also estimate recharge.
Gauging and sampling of piezometers
to determine groundwater movement and
changes in groundwater quality.
Hydrochemistry Regional reconnaissance stream surveys but Hydrochemistry of the area determined.
• Low-flow stream surveys mainly during feasibility studies. Weekly or monthly hydrochemical sampling
of major stream(s) to determine
temporal changes in chemistry of
groundwater.

134 J. S. KUMA ETAL

Fig. 5. Proposed flow chart of the hydrogeological framework for assessing environmental impacts of large-scale
gold mines.

gathered at the same time and consciously built hydrogeological information can be integrated
into a database for later interpretation. with the EIA document. This actually eliminates
The fact that hydrogeological data are available the tendency of looking at time as a constraint and,
at such an early stage means that it can be included therefore, precludes a half-hearted (or absent)
in the EIA report when presented to the EPA for hydrogeological section in the report. This
evaluation before a lease for mining is granted. procedure may also be employed in areas where
Interpretation of these data could be a useful input 'incomplete' EIA reports are a common
to dewatering design and groundwater manage- occurrence.
ment. Any water borehole drilled by a mine
should be seen as an opportunity for carrying out References
test pumping so that aquifer properties can be
determined. Piezometers installed for water ACHEAMPONG, P.K. 1986. Evaluation of potential
evapotranspiration methods for Ghana. Geo Journal,
monitoring need to be kept in working order and 12,409-415.
measurements taken regularly for valuable data to AMEGBEY, N. 1996. West Coast Explosives (Ltd) EIA on
be available for later assessment. the Bulk Emulsion Explosive Plant at Nsuta-Was-
The onus is, therefore, on the regulation agency saw. West Coast Explosive Ltd., Tarkwa, Ghana.
to make it a policy for all prospective mines to ANON. 1980. Low Flow studies. Institute of Hydrology,
include vital hydrogeological information in their Wallingford, Research Report, 1.
EIAs to enhance knowledge of the hydrogeology ANON. 1994. In: LAING, E. Ghana Environmental
Action Plan (Vol. 2), Technical Background Papers
of the area in question. The argument for this by the Six Working Groups. Environmental
procedure is that once the decision is made for Protection Council (EPC), Ghana.
detailed drilling to commence, it implicitly means ANON. 1997. Drilling of Production Boreholes and
that the prospect is of interest and that it also Monitor Boreholes at the Tark\va Mine. Internal
warrants a hydrogeological assessment. The SRK Report, 229708.

HYDROLOGICAL FRAMEWORK FOR GOLD MINE ASSESSMENT 135

APPELO, C.A.J. & POSTMA, D. 1993. Geochemistry, HOUNSLOW, A.W. 1985. Strategy for subsurface
Goundwater and Pollution. A.A. Balkema, Rotter- characterisation research. In: WARD, C.H., GIGER,
dam. W. & MCCARTY, PL. (eds) Ground Water Quality.
BEVANS, H.E. 1986. Estimating stream-aquifer John Wiley, New York, 356-369.
interactions in coal areas of eastern Kansas by HOUNSLOW, A.W. Water Quality Data: Analysis and
using streamflow records. In: SUBITZKY, S. (ed.) Interpretation CRC Lewis, New York.
Selected Papers in the Hydrologic Sciences, US JUNNER, N.R., HIRST, T. & SERVICE, H. 1942. The
Geological Survey, Water Supply Paper. Vol. 2290, Tarkwa Goldfield, Gold Coast Geological Survey,
51-64. Memoir, 6, 75.
BOWELL, RJ. 2002. The hydrogeochemical dynamics KESSE, G.O. The Mineral and Rock Resources of
of mine pit lakes. In: YOUNGER, P.L. & ROBINS, N.S. Ghana; Balkema, Rotterdam.
(eds) Mine Water Hydrologogy and Geochemistry, KOLM, K.E. 1996. Conceptualization and characteriz-
Special Publications. Geological Society, London ation of ground water systems using Geographic
Vol. 198,59-185. Information Systems. Engineering. Geology, 42,
CHIDLEY, T.R.E. 1981. Assessment of groundwater 111-118.
recharge. In: LLOYD, J.W. (ed.) Case Studies in KUMA, J.S. & YOUNGER, P.L. 2001. Pedological
Ground Water Resources Evaluation. Butler & Characteristics Related to Ground Water Occur-
Tanner, London, 133-149. rence in the Tarkwa Area, Ghana. Journal of
DAVIS, S.N & DEWEIST, R.J.M. 1991. Hydrogeology, African Earth Sciences, 33, 363-376.
reprint edn. Krieger, Florida. LERNER, D.N. 1990. Techniques (Part III). In: LERNER,
DICKSON, K.B. & B E N N E H , G. 1988. A New D.N., ISSAR, A.S. & SIMMERS, I. (eds) Ground-
Geography of Ghana. Longman, London. water Recharge; A Guide to Understanding and
DINGMAN, S.L. 1994. Physical Hydrology. Prentice- Estimating Natural Recharge, International Associ-
Hall, Englewood Cliffs, New Jersey. ation of Hydrogeologists. Vol. 8, 99-228.
DUNNE, T. & LEOPOLD, L.B. 1998. Water in Envi- MAU, D.P. & WINTER, T.C. 1997. Estimating ground
ronmental Planning. W.H. Freeman, New York. water recharge from streamflow hydrographs for a
DZIGBODI-ADZIMAH, K. 1996. Environmental con- small mountain watershed in a temperate humid
cerns of Ghana's gold booms: past present and climate, New Hampshire, USA. Ground Water, 35,
future. Ghana Mining Journal, 2, 21–26. 291–304.
EDMUNDS, W.M. 1981. Hydrogeochemical investi-
MEYBOOM, P. 1961. Estimating groundwater recharge
gations. In: LLOYD, J.W (ed.) Case Studies in
from stream hydrographs. Journal of Geophysical
Ground Water Resources Evaluation. Butler &
Research, 66, 1203-1214.
Tanner, London, 87–111.
EIJPE, R. & WEBER, K.J. 1971. Mini permeameter PETTYJOHN, W.A. 1985. Regional approach to ground
for consolidated rock and unconsolidated sand. water investigations. In: WARD, C.H., GIGER, W. &
American Association of Petroleum Geologists MCCARTY, P.L. (eds) Ground Water Quality. John
Bulletin, 55, 307. Wiley, New York, 402-417.
EISENLOHR, B.N. 1992. Conflicting evidence on the PETTYJOHN, W.A. & HENNING, R. 1979. Preliminary
timing of mesothermal and paleoplacer gold Estimate of Ground Water Recharge Rates Related
mineralisation in Early Proterozoic rocks from Streamflow and Water Quality in Ohio. Water
southwest Ghana, West Africa. Mineralium Depos- Resources Center, Ohio State University, Project
ita, 27, 23-29. Completion Report, Vol. 552, 333.
Fetter, C.W. 1994. Applied Hydrogeology. Prentice- RORABAUGH, M.I. 1964. Estimating changes in bank
Hall, Upper Saddle River, New Jersey. storage and ground water contribution to stream-
Grindley, J. 1970. Estimation and Mapping of flow. International Association of Science &
Evaporation. IASH publication No. 92. World Hydrology, Publication, 63, 432-441.
Water Balance, I, 200-213. RUTLEDGE, A.T. 1992. Methods of using streamflow
HEAD, K.H. Manual of Soil laboratory testing. Soil records for estimating total and effective recharge
Classification and Compaction Tests', Pentech in the Appalachian Valley and Ridge, Piedmont,
Press, London, Volume 1. and Blue Ridge physiographic provinces. In:
HEDIN, R.S., NAIRN, R.W. & KLEINMANN, R.L.P. HOTCHKISS, W.R. & JOHNSON, A.I. (eds) Regional
1994. Passive Treatment of Coal Mine Drainage, Aquifer Systems of the United States, Aquifers of
US Bureau of Mines Information Circular, Vol. the Southern and Eastern States, American Water
9389, 35. Research Association Monograph Series. Vol. 17,
HEM, J.D. Study and Interpretation of the Chemical 59-73.
Characteristics of Natural Water, United States SHARP, J.M. & MCBRIDE, E.F. 1989. Sedimentary
Geological Survey, Water Supply Paper, Vol. 2254, petrology - a guide to paleohydrogeologic
263. analyses, example of sandstones from the north-
HIRDES, W. & NUNOO, B. 1994. The Proterozoic west Gulf of Mexico. Journal of Hydrology, 108,
paleoplacers at Tarkwa Gold Mine, SW Ghana: 367-386.
sedimentology, mineralogy and precise age dating SHAW, E.M. Hydrology in Practice, Chapman and
of the main reef and west reef, and bearing of the Hall, London.
investigations on source area aspects. BGR Geol Jb, SHEVENELL, L. 2000. Analytical method for predicting
D100, 247–311. filling rates of mining pit lakes: example from

136 J. S. KUMA ETAL.

the Getchell Mine, Nevada. Mining Engineering, Single Resource, United States Geological Survey,
52, 53-60. Circular, Vol. 1139, 79.
SLOTO, R.A. & GROUSE, M.Y. 1996. HYSEP: A WOODALL, R. 1984. Success in mineral exploration: a
Computer Program for Streamflow Hydrograph matter of confidence. Geoscience Canada, 11,
Separation and Analysis, United States Geological 41-46.
Survey, Water Resources Investigations Report, YOUNGER, P.L. 1992. The hydrogeological use of thin
Vol. 96-4040, 46. sections: inexpensive estimates of groundwater
STONE, W.J. Hydrogeology in Practice. A Guide to flow and transport parameters. Quarterly Journal of
Characterizing Ground-water Systems, Prentice- Engineering Geology, 25, 159–164.
YOUNGER, P.L. 1993. Simple generalized methods for
Hall, Englewood Cliffs, New Jersey.
estimating aquifer storage parameters. Quarterly
THORNTHWAITE, C.W. & MATHER, J.R. 1957. Instruc- Journal of Engineering Geology, 26, 127-135.
tions and Tables for Computing Potential Evapo- YOUNGER, P.L. 2001. Mine water pollution in Scot-
transpiration and the Water Balance publications land; nature, extent and preventive strategies.
in Climatology, (3) Centerton, New Jersey Vol. 5, Science of the Total Environment, 265, 309-326.
185-311. YOUNGER, P.L. & ADAMS, R. 1999. Predicting Mine
WHITELAW, O.A.L. 1929. Geological and Mining Water Rebound; Environment Agency, R&D
Features of the Tarkwa-Aboso Goldfield, Gold Technical Report, W179.
Coast Geological Survey, Memoir, Vol. 1, 46. YOUNGER, PL., BANWART, S.A. & HEDIN, R.S. 2002.
WINTER, T.C., HARVEY, J.W., FRANKE, O.L. & ALLEY, Mine Water: Hydrology, Pollution, Remediation.
W.M. 1998. Ground Water and Surface Water: A Kluwer, Dordrecht.

Predicting mineral weathering rates at field scale for mine water risk
assessment
STEVEN A. BANWART, KATHERINE A. EVANS & STEPHANIE CROXFORD
Groundwater Protection and Restoration Group, Department of Civil and Structural
Engineering, University of Sheffield, Sheffield SI 3JD, UK
(e-mail: s.a.banwart@sheffield.ac.uk)

Abstract: A general challenge to the environmental management of mine sites is the
relatively high costs of site investigation and the associated development of conceptual site
models and parameterization of reactive transport models. A particular problem is the
inability to predict at field scale the rates of processes that give rise to long-term dissolved
contamination due to active sulphide mineral weathering. Mineral weathering rates
determined from laboratory and field observations generally do not agree, often exhibiting a
discrepancy of two-three orders of magnitude.
Recent work on mine waste deposits has demonstrated that this discrepancy can be
explained by considering a small number of bulk physical and chemical properties of mine
rock at field sites. The apparent decrease in mass-normalized rates between bench-scale
batch reactors and pilot-scale column reactors is predicted by accounting for differences in
temperature, where lower temperatures in the column reactors reduced reaction rates due to
activation energy effects. The columns also contain a significantly greater mass fraction of
larger particles that have lower specific surface area and, thus, exhibit lower weathering
rates.
The further apparent decrease in rates between the column reactors and field scale is
predicted by additionally accounting for the spatial variability of sulphide-bearing rock at
the site, which gives rise to only localized weathering. Localized zones of sulphide
weathering are also associated with locally active weathering of silicate minerals due to
lower pH. Hydrological factors are also important due to preferential flow within the field
site, whereby a fraction of dissolved weathering products are retained within immobile water
and do not reach the effluent stream, where ion mass flows resulting from weathering
reactions are determined. These results suggest that application of compiled laboratory data
to prediction of weathering rates at mine sites may be feasible. This is potentially valuable
for application to Tier 1 risk assessment of mine sites, where reliable prediction of
weathering rates from tabulated laboratory data would provide significant information to
support the generally sparse datasets that are available, particularly for orphan mine sites.

Mining activities worldwide contribute signifi- for owners, operators and investors to have a
cantly to the solute loads of receiving streams clear assessment of potential environmental
and aquifers. The contribution of sulphide liabilities on their balance sheets,
mineral weathering associated with mine sites In the following sections we first present an
to sulphate ion loads is currently estimated at introduction to mine water risk assessment
12% of the global fluvial sulphate flux to the methods and the problems associated with
world's oceans (Nordstrom & Southam, 1997). predicting the rates of chemical weathering at
Associated with this weathering flux are mine sites for a range of minerals that give rise to
dissolved metal ions and acidity as potential contamination, and that also help neutralize
environmental hazards. Assessment of current acidity and thus attenuate contamination loads,
environmental risk associated with developed The second section outlines hydrochemical
mine sites is necessary in order to make methods for assessing rates of mineral weathering
appropriate management decisions for monitor- in the laboratory and in the field, and outlines
ing of site emissions, and, if necessary, treatment possible sources for the significant discrepancy
or restoration of environmental quality. Assess- between rates that is generally observed at the two
ment of future risk is critical when developing different physical scales. The third section
licensing plans for new mining activity in order presents results from a site-specific study where

From: YOUNGER, P.L. & ROBINS, N.S. (eds) 2002. Mine Water Hydrogeology and Geochemistry. 137
Geological Society, London, Special Publications, 198, 137-157. 0305-8719/027$ 15.00
© The Geological Society of London 2002.

138 S. A. BANWART ETAL

this discrepancy was successfully explained by addition, there are exceptional commercial
accounting for differences in a few bulk pressures on current or potentially new mining
physical–chemical parameters between the two activity. In both cases, this means that decisions
scales. These parameters are temperature, miner- on environmental liability and risk must be made
alogical composition, ratio of mobile to immobile with the absolute minimum of cost.
water content, particle size distribution and An important consequence of minimal invest-
related physical surface area, and differences in ment in risk assessment is relatively high
pore water pH. The concluding section critically uncertainty in the site conceptual model and
discusses the potential to extend this method- any associated quantitative estimates of environ-
ology for predicting weathering rates at field scale mental risk. This uncertainty results in con-
to other sites and other environmental problems servative decisions on further action. If serious
where mineral weathering kinetics play an contamination is thought to be present, one
important role. consequence may be that inordinately exhaustive
and, therefore, expensive restoration schemes
may be commissioned in order to make sure that
Mine water pollution and environmental risk any future potential liability is removed. A more
Within in the UK, implementation of Part IIA of likely outcome of this uncertainty is that many
the Environmental Act of 1990 now requires a orphan sites will not be redeveloped and further
risk-based approach to the assessment of mining at potentially lucrative sites may be
potential harm or damage to receiving waters, unnecessarily curtailed.
arising from contaminated land. Implications of Figure 1 illustrates a general conceptual model
the new legislation for mines and quarries has for sources of mine water pollution, transport
recently been outlined by the Environment pathways for soluble contaminants and poten-
Agency of England and Wales (Bone 2001). tially sensitive receiving waters that pose targets
The long history of mining activities, from the for contaminant risk. Within this source-path-
Bronze Age forward, has resulted in a large num- way-target risk assessment framework, there is a
ber of now abandoned sites that represent poten- common conceptual approach to mine sites that
tial environmental risk. These orphan sites have arises from considering the oxidative weathering
no legally responsible party attached to them. of sulphide minerals to release soluble and,
Assessment for potential site re-use, including therefore, mobile metal ions and acidity to the
any necessary restoration work, requires either environment. Weathering of the contaminant
public money or private investment that is to be source minerals pyrite and sphalerite are shown
recouped with the redevelopment of the site. In here as examples:

Fig. 1. A schematic representation of contaminant sources, transport pathways and sensitive risk targets in mining
environments. Attenuation processes occur during transport from source areas to targets, thus reducing
contamination levels and risk. The Tier 1 risk assessment methodology presented in Table 1 considers long-term
juvenile contamination arising from active weathering of sulphide minerals above the water table where O2(g)
ingress occurs.

PREDICTING WEATHERING RATES 139

pyrite weathering Ban wart 2002). Key factors affecting the relative
risk from a site include:
• the extent of unsaturated flow processes
where atmospheric O2(g) ingress occurs;
• the extent to which sulphide minerals are
present within disaggregated, crushed or
milled particles allowing greater reactive
sphalerite weathering surface area for reaction;
• the proximity of oxic source zones for
sulphide mineral weathering to discharge
Further consideration is given to the fact that points whereby shorter flow paths and,
solubility-limiting oxide and hydroxide mineral therefore, lower residence times provide
less opportunity for attenuation processes
phases, such as Zn(OH)2(s), are generally much
more soluble at low pH. Acidity release due to (see Fig. 1) to occur.
pyrite weathering poses an environmental hazard
itself. However, it also can lead to a drop in pH,
which generally increases the solubility and, Risk-based methods for assessing mine water
thus, transport of hazardous metal contaminants pollution
to potentially sensitive targets such as aquifers or
receiving streams. Figure 3 shows a schematic representation of the
A particular distinction is made in Fig. 1 'tiered risk assessment methodology'. The
between vestigial contamination and juvenile diagram is based on the 'risk-based corrective
contamination. Vestigial contamination accumu- action' (RBCA) guidelines published by the
lates as secondary mineral precipitates of the ASTM (1995) for petroleum hydrocarbon
metal ions and sulphate that arise from sulphide pollution in the subsurface. The aim of such
mineral weathering in dewatered void spaces in methodology is to guide decisions on the use of
mine environments. A particular relatively short- risk assessment and site investigation resources.
term problem arises when subsurface workings Tier 1 represents the preliminary assessment of a
are flooded, causing these secondary precipitates site. If the site is relatively easy to understand
to dissolve. The time scale for attenuation of the
resulting contamination is controlled by the
hydraulic flushing time of the mine.
Juvenile contamination arises due to ongoing
release of soluble contaminants due to active
sulphide mineral weathering that is driven by
O2(g) ingress above the water table. Contami-
nation is subsequently transported with subsur-
face flow to potential receptors such as receiving
streams. Juvenile weathering thus represents a
relatively much longer-lived contamination
problem. The longevity of contamination is
controlled by the rate of sulphide mineral
weathering and the mass of reactive minerals in
the zone of active weathering. The risk assess-
ment framework described below is relevant to
the long-term risk posed by juvenile contami-
nation where mineral weathering rates, but not
hydraulic flushing times, are the critical control
on the contamination source strength and long-
evity. Vestigial contamination and the linked Fig. 2. Qualitative representation of risk arising from
hydrogeological problem of mine water rebound the oxidative weathering of pyrite and other sulphide
are reviewed in detail elsewhere (Younger et al minerals (Younger & Banwart 2002). Ingress of O2(g)
is enhanced by shallow source zones and those above
2002). the water table in proximity to the atmosphere.
Figure 2 shows this general conceptual model Proximity of source areas to discharge points also
for the relative risk from long-term acidity increases contamination through shorter transport
generation to receiving waters from various paths and residence times, thus providing less
subsurface mine environments (Younger & opportunity for attenuation processes to occur.

140 S. A. BANWART ET AL.

Fig. 3. A diagram of the Tiered risk assessment framework' based upon 'risk based corrective action' for
subsurface petroleum hydrocarbon contamination (ASTM 1995). Tier 1 represents initial site assessment. Sound
understanding of site contamination and risk scenarios allows robust decisions to be made regarding site use and/
or remediation at lower Tiers. Poor understanding can result in a 'tier upgrade', which entails greater effort and
expense for analysis of site data, and often further site investigation.

based on existing information, then a reliable the site data that are required for such
decision may be taken to use the site in its present preliminary assessment.
state or invest in site remediation. If the nature Younger (2000) recently provided a decision
and type of data available are unsuitable for flow sheet for preliminary assessment of the
making such decisions, i.e. the conceptual model longevity and future water quality (relative
for the site is too uncertain, then a decision can acidity and dissolved iron load) of discharges
be made to abandon the effort or upgrade to a from flooded coal mine workings. The strength
higher tier. Additional site investigation, as well of the approach is that it relies on a very general
as data analysis, generally accompany a tier conceptual model and, therefore, requires very
upgrade in order to improve the basis for making little site-specific information. The model was
the best management decision. calibrated against data from a large number of
Each tier is assigned appropriate 'data quality UK coal mine discharges and the resulting
objectives', i.e. the type and quality of infor- uncertainty in the model predictions is
mation required to carry out the particular level quantified.
of assessment and reach a decision. Banwart & A common philosophy that is evident in these
Malmstrom (2001) recently proposed Tier 1 two approaches is the acknowledgement that site
data quality objectives' and assessment method- data are extremely precious. Orphan sites can be
ology for risk-based assessment of mine water expected to have very sparse site data and in
discharges. The methodology is particularly many cases few mine records. Currently, operat-
suitable for spoil deposits and shallow, relatively ing or potentially new sites will probably have a
well-characterized mine environments. Table 1 very limited budget to carry out new site inves-
summarizes the proposed analysis method and tigations targeted specifically at environmental

PREDICTING WEATHERING RATES 141

Table 1. Data quality objectives and site assessment methodology appropriate for Tier 1 (Banwart & Malmstrom
2001).

Assessment step Data required Corresponding data sources
1. Identify 1. Discharge composition 1. Water quality analyses
• Possible pollutants 2. Mine petrology and mineralogy 2. Mine records and geological memoirs
• Source areas 3. Stratigraphy/structural geology 3. Mine records and geological memoirs
• Transport pathways 4. Groundwater flow paths 4. Mine records and geological maps
• Sensitive targets 5. Location of site boundaries, 5. Site maps and plans,
discharge points, aquifers, regulatory authority records
streams, boreholes, etc.
2. Site water balance 1. Recharge to subsurface 1. Hydrological atlas
2. Evapotranspiration 2. Hydrological atlas and knowledge
of mine ventilation
3. Number of site discharges 3. Site maps and plans
4. Discharge flow rates 4. Pumping records and environmental
monitoring records
5. Groundwater flows 5. Borehole monitoring records,
geological maps,
hydrogeology handbooks
3. Discharge solute 1. Discharge flow rates 1. Environmental monitoring records
mass flows 2. Discharge solute concentrations 2. Environmental monitoring records
3. Temporal variability in same 3. Environmental monitoring records
4. Source mineral 1. Identify source minerals at site 1. Geology reports and mine records
weathering rates 2. Discharge solute mass flows 2. Step 2. Results
3. Weathering reaction stoichiometry 3. Mineralogy handbook
5. Source mineral 1. Spatial extent and porosity 1. Site maps and mine
abundance of workings and deposits plans
2. Volume and mass of 2. Site maps and mine
reactive rock plans
3. Mineral composition of rock 3. Mine records, geological memoirs,
mineralogy handbooks
6. Source mineral 1. Source mineral weathering rates 1. Results from Step 4
longevity 2. Total reactive source mineral 2. Results from Step 5
abundance
7. Exposure scenario 1. Longevity of acidity generating 1. Results from Step 6
evaluation minerals
2. Longevity of metals generating 2. Results from Step 6
minerals
3. Longevity of alkalinity generating 3. Results from Step 6
minerals

assessment. Even for sites with extensive mining immediate benefits. First, this reduces the amount
records and a good understanding of the exist- of effort spent in developing a conceptual site
ing or future mining operations, it can be difficult model. Second, by starting from a proven
for owners and operators to meet data quality conceptual model, effort can then be targeted
objectives beyond those for Tier 1 assessment. towards reducing uncertainty in the site-specific
This is largely because of the specialized know- conceptual model and improving quantitative
ledge and data that are required for quantify- analysis of site data.
ing water flow, geochemical reaction rates and There is also a third benefit that is explicitly
contaminant transport pathways in complex linked to current research programmes in the area
hydrogeological systems. of mine water pollution. By developing more
An important strategy to reduce the cost and general understanding of the processes that give
effort involved in Tier 1 risk assessment is to rise to contamination, researchers are able to
develop more general conceptual models for such target underpinning science towards develop-
complex hydrogeological systems. This has two ment of more general mathematical descriptions

laboratory rates and these catchment and aquifer studies that demonstrate a scaling parameters. where the subscript 'py' general discrepancy where weathering rates from refers by way of example to the mineral pyrite: field studies are generally two-three orders of magnitude slower than those observed in laboratory studies. Associ. pro. the ability to predict weathering rates This relationship implies that if the values of m at field scale. then the laboratory rates could be used to stitute a major improvement in data quality predict the weathering rate (Rpy. respectively. widely used in the assessment of watershed which in turn depends on both reacting mineral acidification impacts on surface waters (reviewed mass and the corresponding specific surface area. Successful prediction of known. The modelling solution. mineral mass within a mine workings based on 7). quence of these reactions is that reaction rates are derpinning science has established that such data generally proportional to the amount of reacting will vary in a known way with aqueous solution mineral surface area in contact with aqueous temperature and ionic strength. 3). based on compiled laboratory data py (g) and aPy (m2 g -]) are known for a mine and basic site investigation results. steps 4. In practice it may be possible to estimate ling at Tier 1 risk assessment (Table 1. to that of the dissolved metal ion or complex in The associated science must. This rates allows more fundamental and. environment from that within the crystal lattice be tabulated as handbook values. aqueous solution. Fundamental parameters do not the change in the metal ion co-ordination change with site conditions and can. chemical and microbiological Scaling parameters for mineral weathering processes that determine contaminant loads. 13). A ordering of the crystal lattice terminates with relatively familiar example in geochemical appli. mineral-aqueous solution interface. where the mental parameters to site-specific conditions. more generally applicable parameter sets to be Chemical weathering is controlled by the developed for application in reactive transport kinetics of surface chemical reactions. An excellent overview of the theory of codes thus provide general algorithms to scale chemical weathering is provided by Stumm & the fundamental thermodynamic data to account Morgan (1996. 2 and A particular problem that will be addressed in included references) provide additional reviews some detail in the following sections is the lack of silicate mineral weathering and sulphide of a general conceptual model for the chemical mineral weathering. This would also find important application for mine plans and mineralogical assay information parameterization of more advanced reactive on the ore body and surrounding rock. site. such as watershed acidification where for reliably extrapolating laboratory test results mineral weathering plays an important role. These rates are reported in units such as general scaling procedure to predict mineral mol g-1 day . of the physical.142 S. Un. namely modelling. mass of reacting sample and allows laboratory Studies of mineral weathering at field scale rates to subsequently be normalized to the react- and associated application of weathering rates ing surface area (rlab.1 ) at objectives for contaminant source term model. However. ized to the mass of the reacting mineral or rock ated with this problem is a complete lack of a sample. m (g).1 . of course. mol d a y . The reaction occurs at the vide quantitative methods to scale such funda. m2 g-1) relates the physical surface area to the kinetic data from laboratory studies. chap. by Schnoor 1990). would con. With respect to mine water pollution. therefore. In this case determined in laboratory studies have been surface area becomes the scaling parameter. An important conse- modelling of aqueous chemical speciation. BANWART ET AL. to the complexity and physical scale of mine . some metal bonds maintained to the lattice. mol m-2 day . White & Petersen (1990) reviewed the weathering rates at field scale from laboratory role of the reacting surface area as a scaling data would also constitute a major step forward parameter for mineral weathering rates in detail. in spite of the mineral mass. the site. weathering of minerals that act as sources of Laboratory derived rates are usually normal- soluble contamination (source minerals). White & Brantley for these site-specific effects. The scaling parameter is therefore weathering rates at mine sites. The specific surface area extensive compilations of detailed weathering (a. and cations is the use of general thermodynamic others to aqueous species such as chemically databases with computer codes for equilibrium bound water molecules. in the understanding of other environmental Although there is currently no general method problems. chap. transport models corresponding to Tier 2 risk values for specific surface are generally not assessment (Fig. therefore. A. (1995) and Younger et al (2002.1 ). White & Petersen (1990) Equation (3) shows the relationship between reviewed a number of datasets arising from field weathering rate.

g.1 ) yielding a value with units of m2 g-1. must be corrected for constitutes 'reactive' surface area (Ar. and therefore on the applying such methods to analysis of field data.1 ) the amount of pyrite in the rock.l1d 4. Fs.3). White & Petersen (1990) implies that solute mass flows originating from a also present considerable discussion on what single mineral.6 Biotite 8% Mg22+. The value for Ar is obtained dissolved: by dividing the weathering rate at the site.1% Cu 5.6% so 4 2+ 1. The range of values As explained below. m 2 g . Equation (4) relates reactive surface area.1 ) is treated as a direct sites. and solute fluxes from another site where This approach assumes. (mol K g .9 Anorthite 6% Ca2+ 8. in considerable detail. This 'calibrated' value could then be used to scale Values for Ar such as those in Table 2.8 Pyrite 0. not accumulated within the work- similar geology and mining history will have a ings as secondary precipitates or in immobile common conceptual description and similar site water.1 X 10" 0. e From Stromberg & Banwart 1994. the mass fraction of pyrite in the rock.12 l1d Albite 13% Na 3. An alternative approach to this scaling problem Sulphur mass flow must also be divided by a is to estimate solute fluxes at the field site.9x 10. Table 13. and how to test them. example. d From Stumm & Morgan (1996.1 day .8 X 10" 12 1. Reactive surface area estimated at field scale Mineral Tracer in Weathering rate from Laboratory rate. e.1 0 e 4. As an example here. This method. this is achieved by spans a five-fold difference. Ar. where 2 mol of dissolved obtain an empirical measure of Ar for rock that is sulphur are released for each mol of pyrite weathering at the site. can be laboratory weathering rates to similar types of used in equation (4) along with tabulated field sites. PREDICTING WEATHERING RATES 143 sites. The simplest and how this may relate to the physical surface approach is to assume that the fraction of reactive area of weathering rock. Albite (Al) and anorthite (An) dissolution rates correspond to weathering of a single plagioclase with a composition 70%A1-30% An. normalized against rock mass (mol g-1 day-1)2.3 X 10~ 13 2 . that sulphate ions are conserved in the ology is based on the premise that sites with discharge. it is still possible in some cases to get order determined from a variety of solutes in the of magnitude estimates of contaminant loadings.1 a Biotite gneiss with mica schist reported in Stromberg & Banwart 1994.1 X 10.5 X 10" 12 3.7 X 10 3.5x 10" l0e 2. b From Banwart et al (1998). and factor of 2 to reflect the weathering stoichiometry compare these with laboratory rates in order to of pyrite (FeS2(s)). This relationship extremely uncertain.l1d 1.%) (molkg .0 Chalcopyrite 0. physical structure of geological material associ. 2.2 X 10~ 11 5 x10 . c Determined by dividing weathering rate from tracer flux in field by the laboratory rate. or formation of sulphate-bearing secondary morphology and size distribution of weathering precipitates is a potential source of error when particles in the rock mass.1 s-1) (m2kg. pH 4 Reactive surface area Abundance in discharge tracer flux in fieldb (mol m-2 s-1 ) at field Scalec Fielda (vol. mol m-2 d a y . limitations of such assumptions. by the laboratory weathering rate (rlab2. ated with a mine workings. pyrite. determined for field site with mineralogy given in column 1. . ') + . The specific surface area of an individual measure of pyrite weathering at field scale. Xpy. this information is laboratory weathering rate. the mass flow of sulphate normalized to total rock Scaling weathering rates between similar field mass. Banwart & Malmstrom (2001) discuss the characteristics. discharge from a field site. There is not a simple surface area corresponding to pyrite is similar to relationship between the two parameters. i. using pyrite as an measurements have been made.e. which is remarkably transferring information about weathering rates small given the uncertainties involved. but it mineral at a field site presumably varies should be borne in mind that sorption of sulphate considerably with the mineral composition. Table 2 shows calibrated values for Ar laboratory data on rpy and estimates of pyrite Table 2. Because of the spatial to the sulphate mass flow at a field site and to the heterogeneity of field sites.

Table 3. Net acid generationb • Reacts crushed rock with hydrogen • Similar objectives as for ABA peroxide to oxidize sulphide minerals • Accounts specifically for acidity due to oxidation • Measures the amount of oxidizable of sulphide minerals sulphur in rock • Does not account for rates of weathering 3. Assessment methodologies for source mineral weathering in mine water risk assessment Assessment method Description of methodology Relation to data quality objectives a 1. Batchexperimentsd • Kinetic test of weathering rates • Provides detailed weathering rate data • Solute accumulation rate related to • Does not account for flushing individual source minerals in mine environments 5. contamination is anticipated rock with hot HC1 with • Does not account for rates of weathering back-titration to determine amount of • Does not react at field conditions HC1 neutralized • Assumes all sulphur is acid-generating • AP determined by relating chemical analysis of total sulphur content of rock to equivalents of H2SO4 2. Acid -base accounting (ABA) • Compares neutralization potential (NP) and • Provides pre-mining assessment of contamination acidity potential (AP) potential • NP determined by reacting crushed • If AP > NP. Humidity cellsexperimentsc • Reacts mill tailings under • Assesses time to exhaustion for hydraulically-unsaturated conditions source minerals • Dry air blown through tailings • Realistic representation of tailings deposit for fixed time followed by conditions • Moist air blown through for • Tells if alkalinity generating minerals same amount of time have greater or less longevity • Then flushed to capture oxidation than sulphide minerals products • Cycle repeated until source minerals are exhausted or pH drops 4. Columnexperimentsc • Solute fluxes measured in column • Provides detailed weathering rate data effluents • Accounts for flushing of reaction • Fluxes related to individual weathering products rates of source minerals • Requires significant lab prowess .

Implications for extrapolating weathering rates from the laboratory to field scale. the corresponding laboratory data are still useful for understanding develop- ment of water quality and its temporal trend. This implies that the factors responsible for the discrepancy are largely independent of the various dissolution mechan- isms for different minerals. The range of values in Table 2 suggest that this scaling discrepancy between Ar and a can be similar for many minerals at a given field site (Stromberg & Banwart 1994). carbonate and alumino- silicate mineral weathering (see Banwart & Malmstrom 2001). A common feature of these studies is that empirical values for reactive surface areas. Determination of weathering rates by aqueous chemical methods There are a number of assessment methodologies for source mineral weathering rates that range from relatively simple laboratory leaching tests to much more sophisticated field-based methods that resolve spatial variability in weathering . are always two- three orders of magnitude lower than the specific surface area determined in laboratory studies (White & Petersen 1990). derived from Ar and mass of rock at the field sites. Although the exact weathering rate for any individual mineral at a field site will probably not be possible to predict from laboratory rates alone. Better site information such as discharge rates. when assessing possible impacts of a planned mining operation for sites with similar geology and hydrology. The main drawback is that even if relative rates of weathering are known. the contamination lifetime cannot be predicted with any certainty. the longevity of source minerals depends on the absolute rates of weathering at a site. Calculations of this type thus help to answer the question 'how bad will the contami- nation be'. geometry of workings and mineralogical information are all important pieces of information to improve estimates of (or to help interpret) contamination loads.PREDICTING WEATHERING RATES 145 content as mass per cent in order to obtain a very crude estimate of solute mass flows in the absence of water quality data. water quality data. In the absence of these rates. Because the development of acidic versus alkaline mine water discharges depends on the relative rates of pyrite. for example. at least the relative rates of weathering for different minerals at a field site may be similar to those determined from laboratory studies.

release with time for a sample of mine waste Two particular types of laboratory methods to rock. concentration data are then used to determine However. carbonate weathering rates that can be extrapolated to the and silicate minerals in relation to pH buffering field. First. which have the advantage of minerals within the sample. i. an important conclusion from the cited work shown in Fig. to establish the extent and rate of acid are restricted in their ability to provide mineral consumption by weathering of oxide. 4b is that particles with Batch reactor methods a nominal diameter greater than approximately Batch reactor experiments are a relatively simple 4mm contribute negligibly to mass of solute approach to determining the potential for a given released by the weathering reactions. rates to the weathering rates of individual Batch reactors. strated by the fact that small particle sizes release are described below. This large mass of rock at a field site. and to relate these release neglected. can be effectively laboratory conditions. to prevent diffusion- measurement results. the problem remains of how to apply ion release rates from the reacting material. In this case. Static tests are based on analysis conditions with typical experimental results of total amount of weathering products in order reported for a temperature of 25°C as a standard to establish chemical capacities for acidity and condition. The batch reactor solutions are achieved. Figure 4a shows a drawing of a typical water to rock ratio within a batch reactor is far reactor configuration. This leads to problems with gas access weight of solid sample and a known volume of and the concentration of pore water solutes aqueous solution. interpretation of pH can give a general indication of the reactor the data is generally fraught with difficulty due to performance and sample reactivity.e. Periodic the complexity of real environments. higher than it would be in any natural rock The reactors are usually set up using a known system. that might be approached for flushed known composition to maintain fixed O2(aq) voids in subsurface mine environments. Secondly. namely batch and column experiments. Oxic or anoxic conditions can be geological formations. BANWART ET AL rates. used to obtain the data for that study.146 S. general scaling procedure is presented and tested Figure 4b shows a time series for copper ion whereby laboratory rates may predict field rates. If such laboratory data to reliably predict what the solutes can be reliably related to the source corresponding rate of weathering might be at a minerals present. is not concentrations. Cu release rates can be related determine mineral weathering rates have been to the source mineral chalcopyrite (CuFeS2(s)). Alterna. A. . often with the solution ionic with respect to comparisons of calculated strength fixed by a swamping electrolyte of mineral solubility equilibria. the solution controlled by maintaining the reacting solution in composition is constantly changing as a result of equilibrium with a specified gas such as N2(g) to mineral dissolution. These The effect of reacting surface area is demon- methods. Reactors methods 1 and 2) and kinetic tests (Table 3. are often maintained at constant temperature methods 3-7). Time series of solute weathering rates can be reliably determined. ions much more quickly than larger particle sizes. These methods can be controlled mass transfer in the fluid film outlined in broad terms as static tests (Table 3. then weathering reaction field site.e. alkalinity production. can be left to run over a period of months. This is for two main reasons. the reactions. i. so that steady-state con- maintain anoxia and N2(g)-O2(g) mixtures of ditions. much greater known composition in order to avoid changes in dilution of released ions occurs in batch reactors speciation due to changes in ionic strength during than would normally occur in pore waters of the experiment. a potentially rates for source minerals can be determined. i. surrounding the reacting solid sample. Although field methods capture processes that Continuous monitoring of the temperature and are actually occurring at a site. while kinetic tests follow Batch reactors can be set up as a large number reactants and/or products with time in order to of relatively low-maintenance experiments that obtain reaction rates. This greatly geological sample to generate contamination simplifies a possible scaling procedure as a under hydraulically saturated conditions. It can also be used being a relatively simple laboratory procedure. Table 3 outlines these methods and the stirred sufficiently rapidly in order to maintain assumptions involved in interpreting the well-mixed conditions.e. laboratory tests are carried out under solution composition allows the reaction pro- controlled conditions where source mineral gress to be followed. In the subsequent section. In fact. above the method can also be applied in order to ascertain 'particle-size cut-off of approximately 4mm release rates of contamination under controlled (specific to the sited study). sampling and chemical analysis of the sample tively.

S. Raes are evaluated by rate of solute accumulation in solution with time. 4. (b) Batch reactor data showing copper ion release from the weathering of chalcopyrite. while larger size fractions show negligible extent and rate of weathering (Stromberg & Banwart 1999a).FeS2(s)).1-S. PREDICTING WEATHERING RATES 147 Column reactor methods enough. . If they are run for long Fig. Rapid and extensive copper ion release results from the smaller size fractions. In addition. they may reach steady-state conditions.-bearing mine rock.3 refer to three replicate rock samples and dp refers to the nominal diameter denoting particle size. thus offer the opportunity for realistic water to column experiments provide the opportunity to rock ratios and gas access. (Cu. (a) A schematic diagram of a batch reactor for assessing ion release and the related rates of mineral weathering in geological solids. where the percolating solution evolves steadily Column experiments may be run under hydrau- from the chosen influent to an effluent of lically saturated or unsaturated conditions. and temporally constant composition.

which. Typical runs composition to preserve its composition.g. The reservoir holds a source of influent. mine water. V). as a 1 m column can provide a effect of varying residence time on effluent reasonable analogue of the top metre of spoil in a chemistries. A. with air trapping. glass sinter or porcelain) than the with control of release by equilibrium with underlying material so that the disk will transmit primary or secondary minerals. although they may be present in quantities too . etc. column reactor. influent (< 1 cm3) Teflon flow cells in which pH. i. It may be necessary to pressurize sandstone and coal. This includes assessing whether (redox potential. 5a): reservoir. as it was felt that multiple sampling heap but may be quite incapable of providing ports at intervals down the columns would induce information on greater depths. sample prep. designed to extract carbonate buffer system. Sulphur concen- water only when saturated. 5b Determination of weathering rates from were obtained from a reactor with the following experimental data. consistent in this range of flow suction (e. the column itself and sample collec. groundwater). The columns were filled with material crushed otherwise the influent may be either a natural to a median grain size of 1 mm. Unsaturated columns element release. via in-line flow providing valuable information on weathering cells to measure parameters such as pH and Eh mechanisms. while unsaturated columns Figure 5b illustrates steady state results for require irrigation by downwards flow driven by sulphur and iron release from a 50cm long gravity plus optional enhancement by additional column containing approximately 7kg of Coal pumps. incorporated hydrological and geochemical 100% N2 to preserve anoxia. Variation of critical as heterogeneities can lead to preferential element concentrations as a function of flow flow within columns and concomitant problems rate may be used to infer mechanisms for with data interpretation. Column lengths were varied to investigate the larly challenging.1 . if required. and dissolved oxygen content were continuously tion. Sample collec. Release rates may be although common elements include the need to combined with bulk and mineralogical compo- prevent flow down the sidewalls. and the controlled purely by chemical kinetics (solute question of averaging element fluxes over flux is independent of flow rate) or whether periods that may be considerable at low flow control of ion release by diffusion-controlled rates must be addressed. into a receiving solubility equilibria between effluent and source vessel. tion involves the flow of effluent. fix PO2(g) for specific information regarding water and gas distribu- oxic conditions. 1 8 m l m i n . (inductively coupled plasma-atomic emission In a saturated column it is advantageous for the spectrometry) for cations and by Dionex ion flow to be pumped upwards to avoid problems chromatography for anions. controlled reactions. influent distribution and effluent collection. the possibility sition data to calculate dissolution rates of the for sampling ports along the length of the minerals concerned. This last is particu. and element release rates. preferential flow. were run in the columns for the influent reservoir with a gas of known periods of approximately 3 months. Influent was deionized water There are four essential components to any and effluent passed through low-volume column experiment (Fig. and the problem of representative and with the identity of secondary minerals. BANWART ET AL study element release as a function of flow rate. Influent application is tion. aration and length scaling. The associated results in Fig. ranging from lead and zinc ore- made up in the laboratory (e. Unsaturated columns were con- conditions. reproducible packing procedures. representative with high-suction glass sinters top and bottom for sampling for column materials. although an tively independent of flow rate up to alternative option is to use a disc of higher matrix 0 .g. mass transfer limitation occurs (solute flux Figure 5 a shows a schematic diagram of a increases less than proportionally to flow rate). The details of column design depend entirely indicative of element release by kinetically on the desired outcome of the experiment. synthetic bearing limestone to Coal Measures mudstone. Care must be proportional to flow rate). whether ion release is taken to avoid evaporation at this stage.148 S. is still complicated by issues such as structed from 10cm diameter uPVC drainpipe heterogeneity within the columns. Samples were generally achieved using a peristaltic pump to taken every weekday and analysed by ICP-AES feed the column. fix PCO2(g) for the characterization procedures. the simplest case this may be deionized water. Iron concentrations are effec- frequently use sprinkler systems. although care must be taken column. In monitored. Homogeneous distribution of influent is Measures sandstone material. A variety of solution (e. rainwater) or solutions materials.e. Eh application. which may be a simple flask or an auto- or secondary minerals occurs (solute flux is mated fraction collection system. tration is inversely proportional to flow rate. and extrapolation to field specifications.g.

are determined from effluent discharge rates and the ionic composition of the effluent. (b) Effluent sulphate ion concentration (•) as dissolved sulphur and dissolved iron concentration (O) are plotted against flow rate for a hydraulically unsaturated column. which is consistent with thermodynamic solubility equilibrium control. Kinetic control of pyrite mineral dissolution results in the observed inversely proportional dependence of sulphur (S) concentration on flow rate. Iron concentrations are independent of flow rate in the range flow rate <0.18mlmin . (a) A schematic diagram of a column reactor for assessing ion release and the related rates of mineral weathering in geological solids. In this case. it is likely that dissolved iron species are maintained in equilibrium with secondary iron oxyhyroxide mineral phases. PREDICTING WEATHERING RATES 149 Fig. 5. . Solute mass flows in the column effluent. and associated weathering rates of the source minerals.1 .

It seems likely that tation along transport pathways from the source these factors will be common across a wide range zone to the point of sampling. Contami- Within the Tier 1 risk assessment methodology nant solutes originating from such 'hot- outlined in Fig. at the time of flow measurement allows particularly below the water table. sites can have a large impact on contami- trol on element concentrations. only be a small fraction of that present. the field to large parts of a site being anoxic and results are also dependent on a number of therefore unreactive. The scale dependence greater specific surface area and thus exhibited by the Aitik data is. the extent of such material may be limited. . significantly faster than those at field sites watershed and aquifer studies suggests that at with low average temperature. mined in the laboratory and the field site. physico-chemical characteristics. Owing to least some of the quantified effects are of gen- the activation energy effects (Equation eral applicability and importance when extra- 10). the field. (2) that solutes can be Elucidating these five factors has resulted in a reliably assigned to the specific source minerals. mineral weathering is faster at higher polating weathering rates from the laboratory to temperatures. which is reviewed below. assumptions such as: (1) the solutes arise solely from mineral weathering. 2. with a weir installed in the effluent channel or by • The availability of O2 may be much more bucket and stopwatch. well-characterized deposits of mining waste ation of laboratory results extremely uncertain. Such data results in some contamination being require measurements of discharge flow rates and retarded in immobile water. lost to sorption or secondary mineral precipi. of mine sites. Grab samples of effluent restricted in the interior of mine sites. If the impact of these factors on weathering rates at field scale can be generally quantified. The fact that Laboratory tests at room temperature are the scale dependence exhibited by the Aitik expected to yield weathering rates that are data is consistent with other laboratory. A. for Scale effects that are commonly cited as which weathering rates have been previously contributing to discrepancies between laboratory published. it would be possible to use the large Resolving the scale dependence of weathering compilation of laboratory data on mineral rates between laboratory and field weathering rates to account for source mineral As discussed above. Material that is suspected to be reactive is usually used for the types of laboratory tests Solute mass flows from field sites outlined above. This means to determine solute concentrations and the that the actual extent of reacting rock may associate solute mass flows in the effluent. BANWART ET AL small for identification by bulk techniques such • Spatial variation in mineralogy at field as X-ray diffraction. rock at the Aitik site in northern Sweden. The cant mass but little reactive surface area magnitude of this discrepancy is similar to for weathering. Flow rates contributing to contamination in the are typically determined by either flow gauging discharge. predictable by quantification of the effects of a • Temperature effects can be significant for few critical and readily available. nant generation in discharges. general conceptual model for weathering pro- and (3) that solutes are conservative and are not cesses in mine sites (Malmstrom et al. The site is suitable as a model and field observations of mineral weathering system for investigating the apparent scale rates are: dependence of mineral weathering rates. Laboratory tests usually those observed for weathering studies in natural focus on smaller particle sizes that have catchments and aquifers. to a large degree. show- ing a two orders of magnitude discrepancy • Particle size effects where large rock between mass-normalized weathering deter- aggregates at field sites contribute signifi. minerals are determined in Step 4 by solute mass • Preferential flow through mine workings flow measurements at a field site.150 S. At field sites. the very large differences in weathering at field sites. due As for batch and column experiments. bulk-averaged field sites in cold weather regions. weathering rates for source spots' may be diluted in the discharge. can exert considerable con. the physical scale and complexity between A large body of data is available from the laboratory tests and field sites makes extrapol. where subsequent chemical analysis to be carried out O2 diffusion is much slower. 2000). rather than the associated solute concentrations. weather more quickly.

5) volume fractions of the minerals. Sweden acidity consuming processes. assuming that Ca2+ The dominant acidity producing processes in the originates from plagioclase and Mg + from Aitik waste rock heaps have been identified to be biotite. 1999a. b. The total solute flows from physico-chemical parameters.%.8-4. which are also dominant silicates in the Aitik waste The Aitik mine. or is no longer reacting sufficiently rapidly to release significant amounts of Ca2+. The Aitik waste rock heaps currently extend over average concentrations of solutes were multi- an area of about 260 ha. weathering processes in the heaps are transported This case study assesses the situation prior to from the site via the ditches. of Cu2+ and sulphate. solutes originating from activities. Hydrological investi- 1994. The surface The equivalent flow of sulphate was subtracted temperature at the site varies considerably. muscovite and K-feldspar. with a depth of 15 -20 m. which is collected in two ditches and used as process water in the enrichment plant. which contain rock material with pyrite. gas ranging from 3 to 21 vol.5Mg1. To estimate the solute mass flows from the site. arising from chalcopyrite weathering was with measured oxygen concentrations in the pore accounted for using the mineral stoichiometry. with an rates at the site. rock that is deposited in heaps at the site (Fig. is Europe's largest proton consumption at the site (Stromberg & operating copper mine. has a pH of 3. Hence. porosity and The weathering of biotite. The low pH of the drainage water indicated that available calcite has been depleted. The Weathering rates for chalcopyrite. PREDICTING WEATHERING RATES 151 Resolving the scale dependence of composition of 30% anorthite. The dominant reaction for copper mobilization at the Aitik site has been identified to be oxidative weathering of chalco- pyrite (Stromberg & Banwart 1994). Rchp. Site description for example. is much slower and contributes less to the northern part of Sweden. The weathering of. respectively. from: & Banwart 1994). producing also waste Banwart 1994). Hence. with from the total in order to obtain the pyrite average winter and summer temperatures of — 15 weathering rate: and 15°C. The reported areal source of Ca2+. the site were obtained by summing flows in chemistry and hydrology (Stromberg & Banwart ditches 1 and 2 (Fig. Therefore. Hence. located close to Gallivare in the rock. CaAl2Si2O8. for assessing protective measures after mining 1992). Rpla. gations and groundwater monitoring indicate that Boliden Mineral AB. Rpyr were determined from the mass flow fragments up to the order of 1 m diameter. The sulphate remain predominantly oxic over their full depth. plied by average flow rates in the ditches draining and have been well characterized in terms of the waste rock heap. 6). NaAlSi3O8) then remain as the major weathering rates at the Aitik mine. when accounting for the mineral oxidative weathering of pyrite and subsequent stoichiometries. The drainage water from the heaps. 6). extent and height of the heap. average water content of about 10 vol. and of plagioclase (with a calculation of the volume of the individual . and 70% albite. Site assessment Silicate weathering rates were calculated from Ca2+ and Mg2+ mass flows. we consider plagioclase as the mineral chemical analysis. Rbio. The owner. we obtained the weathering rates precipitation of ferric oxyhydroxide (Stromberg of biotite. K(Fe1. whereas the temperature in the major part of the heaps was fairly constant around 4°C. solute flows remediation at which point the heaps were in these ditches represent integrated weathering unsaturated with respect to water. respectively.%. Eriksson et al 1997). mineralogy. and plagioclase. and heaps. allowed (AlSi3O10)(OH)2. as has been demonstrated for larger rock Mineral abundance was quantitatively charac- particles at the site (Stromberg & Banwart terized from drill cores by thin section and 1999a). The low pH in the drainage water indicates that available carbonates (mainly calcite) have been depleted to a large degree. has applied dry covers on only a small fraction of the water flow infiltrates the heaps following the outcome of a programme the subsurface below the heaps (Axelsson et al.

no 'cali- bration' of weathering rates is necessary. Rg the universal gas minerals. 1997) as the ratio between mobile and were estimated from solute mass flows in the ditches. Table 5 compares the resulting weathering rates determined from solute release rates in batch reactors and the solute mass flows from the columns and from the field site. they are not the respective weathering rate. when spatially constant and T the absolute temperature at each averaged over the field site are approximately scale: half the value determined for the rock used in the batch and column experiments. with Ea being relative abundance of contaminant generating the activation energy. The Arrenhius equation (equation 10). The column Weathering of both sulphide and silicate experiments demonstrate an approximately one minerals is also known to depend on proton order of magnitude decrease in weathering rates. kB. activity with rates being higher at lower values of compared to the batch experiments. Systematic differences are noted for correction for temperature is quantified using the rock mass. This using the respective mineral densities and factor accounts for weathering that occurs in molecular weights. aF = 0. mesocale column reactors and associated with ion release to solution. The value for the field site is and solute concentrations. and paths and immobile water. 6. which is based on solute mass flows Comparison between laboratory and field in the discharge. temperature and water flow rates.152 S. for temperature converted to the molar amounts for each mineral (T). being considered: . i.65 and the value for the column experiments is ac = 1 (Eriksson et al 1997). The pH effect is accounted for using demonstrates a further approximately one order equation (11). conditions and data Weathering rates are known to depend in a predictable way on temperature based on Table 4 compares the physical and geochemical activation energy effects for the breaking of conditions encountered between bench-scale chemical bonds in the crystal lattice that are batch reactors. The factor a C/F is hydrological and accounts Fig. although obtained by dividing moles of each mineral by solutes are released to solution. transported with the site discharge and so do not contribute to the observed weathering rate at field scale. BANWART ET AL. The field pH. Mineral life times were zones of immobile water where. The different B factors represent a correction of minerals in the deposit. it the locations of the shielding ditches that capture the has been experimentally quantified (Eriksson run-off from the heaps. Diagram of the Aitik mine site showing the for the potential existence of preferential flow extent of the open-pit mine and waste rock heaps. pyrite and chalcopyrite. within the pH range compared to the results from the column reactors. with n being the reaction order of magnitude decrease in weathering rates. with respect to protons.e. Equation (9) thus represents a true prediction and does not rely on a priori information on weathering rates at field scale. pH and particle size distribution (PS). Weathering rates at field scale et al. These volumes were the batch weathering rate. The the field site. Scaling procedure for predicting column and field rates from laboratory rates Equation (9) relates the relative weathering rate in the column or at the field site (RC/F) to the rate determined in the laboratory batch reactors (KB): The scaling factors identified above have been calculated from independent information. total water content in hydrological tests using obtained by simultaneously measuring water flow rates unreactive tracers. A. For the Aitik case.

5 -10.9e -9.5Xl0l0d 0.1 ) (molm .5Fe1. A is total area. e is porosity and ps is density of the solid material.0 -11.5 m2.7Ca0.35 and ps = 2. A = 0.5 a Estimated assuming stoichiometric mineral dissolution with Na+.2 Biotite -9.3Si2.8 -13.5 -11.0 -12. c Stromberg & Banwart (19996). Mineral weathering rates as estimated from observed tracer release rates. chalcopyrite.originating mainly from plagioclase ((Na0.82xl0 3 d 9. C and F refer to the batch.2 s-1) (molm -2 s-') (molm 2 s-1) Pyrite -8.011 0.8 -13. Table 4.7 Chalcopyrite -9.1 -11. Cu2+ and SO24. b Stromberg & Banwart (1999a). .0057 0.c logR Fa. their Table 1) and M is the total waste rock mass (Table 4).0009 0.9 -11.2 0.2 1-4 3.2 –13. in terms of the logarithms of the mass normalized weathering rate R and the associated rate coefficient k = R/a sm . A = 2. c Tracer release rate from Stromberg & Banwart (1999b. Table 5.3 -10. in terms of total waste rock mass M.3 0. ym.9e -10.5 -11. temperature T.b log RC a. biot. chalc. chalcopyrite (CuFeS2).15 0. The labels B.3 .3 . 20-23 3.10.Table 4.0025 0.5 0.1 s .20 ColumnC 1. in the field (8).9 -11. with asm being the specific surface area for mineral m.011 0.15 .1e . in the column experiments (10).6 X 106 m2. d Calculated as: M = HA(1 — e)ps where H is height.7O8). biotite (K(Mg1. Mg2+.7 -10. the average H = 20m. d Tracer release rate estimated as EQ i C i /M where Qi is water flow and Ci is tracer concentration in drainage ditch i of the two drainage ditches at the Aitik site reported by Stromberg & Banwart (1994. Table 4). and pyrite (FeS2) and chalcopyrite. water flow Q. pH and volumetric mineral content.4 Plagioclase -10. average of triplicates).15 0.3)Al1.8 X 103 kgm .3 -12.9 -12. plag denoting the minerals pyrite.3 ) Batchb 0. e Stromberg & Banwart 1994. Characteristics and prevailing conditions for the three observation scales.4 -12. b Tracer release rate from Stromberg & Banwart (1999a. respectively. respectively Rock mass Water flow Temperature pH Mineral content ychal (m3 m .8-4. c Corrected for Chalcopyrite dissolution.2 X 10-9 4-10 3. H = 2m.0025 0.5)AlSi3O10(OH)2).d log kB log kc log kF (molkg-1s -1 ') (molkg-1 s -1 ) (molkg .19 a Plagioclase with a composition corresponding to 70% anorthite and 30% albite.20 Fielde 9.080 0. column and field scale.3 ) ybiot (m3 m 3) yplaga (m3 m -3) M (kg) Q (m3s-1) T(°C) y p y r (m 3 m . respectively Mineral logR Ba.35 and ps = 2.8 X 103 k g m .6 -11. biotite and plagioclase. e = 0. with the index m = pyr. e = 0.

processes that affect and is constant on all scales (Stromberg & the solute mass flows. implies an additional scaling collected from the columns and Aitik mine site factor accounting for differences in mineral do show a small fraction of Cu associated with distribution between the observation scales. the quantification of column and site discharges. batch experiments. bution of the various factors to the observed scale The correction for the particle size effect is dependence. ki were obtained from the dependence of chalcopyrite weathering rates. when normalized to crystal lattice network for silicate minerals. 2000. are rock mass. It with those predicted from the laboratory rates must be noted that sulphide weathering rates. This suggests shown to be more reactive. For both column and field rates. in using equation (9). presumed to have a significant effect on the A further strength of the proposed scaling weathering rates when compared across the three procedure is that it appears to capture essential scales. it is assumed. ym has been determined by point 1999b). site-specific conditions and must then be conceptualized and modelled. to a large degree. which is both sulphide and silicate minerals. However. will generally considering the relatively small number of bulk be microbially mediated. This suggests that additional (Malmstrom et al. results suggest that these factors do. sorption processes at the site. The batch reactors largely comprise rock features of the observed scale dependence for mass from the <4mm size fraction. explain the discrepancy between labora- For the individual minerals. These characteristic of the Aitik site is the very low . the quantified by equation (12) with xi being the largest contributions to the discrepancy arise particle size fraction that is associated with the from temperature and particle size effects. A. used fractions were studied separately (Stromberg & as a tracer for chalcopyrite weathering in the Banwart I999a). Analysis of where ym is the mineral content and as is the total secondary hydrous ferric oxide precipitates specific surface area. The rates are predicted well the pH range described in Table 2. we use activation tory and field observations of weathering rates. 7 show the relative contri- fraction >4mm. is being removed by the specific mineral surface area. energies and reaction orders from the literature Because these factors are common for subsurface (Malmstrom et al 2000.154 S. a particular physico-chemical parameters considered. compared to particle size distribution at the considered scale the other minerals. in which individual size A likely explanation is that dissolved Cu. Fig. 1 in the supporting factors may have a significant impact on the scale information). For this mineral phase (Stromberg & Banwart the Aitik case. yielding primary minerals. aC/Fsm = ymc/Fas. based on previously published results (Mal- weathering rates at mesoscale and field scale mstrom et al. 7 also shows that chalcopyrite The product xiki is summed over the entire is not predicted particularly well. table 3 in the supporting environments. Inspection of Fig. The rates. This scavenged Cu represents a poten- counting using light microscopy (Stromberg & tially significant sink within the site and would Banwart 1994 and included references) and as has introduce a corresponding error to the estimated been determined using the N2 adsorption method weathering rate. as column and field contain progre.1 ). besides the weathering of Banwart I999a. In general. 2000 and included references). BANWART ET AL. The chemical mechanisms of oxidative weathering of column and field should thus. that Figure 7 compares weathering rates for the the weathering processes are rate limited by column experiments and the field investigations abiotic surface chemical weathering reactions. as= l000m 2 kg . as the detailed boulders of 1 m or more in nominal diameter. it is likely that they will play a information). while the column that the procedure is likely to be robust with experiments contain the size fraction <10cm respect to the range of minerals that may be and the field site consists of particle sizes up to considered in other applications. ssively greater mass within the less reactive size The insets to Fig. Moreover. corresponding weathering rate ki: Additional contributions when scaling from column to field are the lower mass fraction of sulphide-bearing rock at the field site and the presence of immobile water within the heaps. similar role for a much wider range of sites than The 'particle-size cut-off' described above is the one considered here. may also prevail under other the scaling factoraC/Fsm/aBsm= YC/F/YBm. and the hydrolysis of the slower weathering rates. show progressively sulphide minerals. For the site-specific quantification of the A comparison of predicted and measured temperature and pH factors. fundamentally different.

Open and filled symbols denote the column and the field scale. such as the ratio of correction for all factors included in equation immobile to mobile water and the particle size (9).C/FpHand the — log beta. "perfect prediction". A – log yC/Fm/yBm and D — log a C/F. 7. Modelled field. where microbially mediated weath. Also. be quite general. Some. temperature and pH factors must then be based However. in addition to knowledge. are empirical and site specific. PREDICTING WEATHERING RATES 155 Fig. in fact. data. factors are observed to vary significantly between The quantification of the effect of preferential sites. annual average temperature that inhibits micro. although immobile. water when determining weathering rates from mination of this effect would thus apply also for solute mass flows at the field site. kB. D chalcopyrite. solutes also contribute to the contamination legacy of the site. listed in Table 5. Application of the observed discrepancy in weathering rates proposed scaling procedure to other sites and between laboratory and field observations. where modelled rates equal observed rates. against the rates that were estimated directly from observed tracer release rates. would still require must be accounted for. The solid line represents the ideal case. It conditions. Although this biotic weathering processes. the re- weathering process and would be the same for sults observed here may. ^ biotite.logBC/FpH. dure is sensitive to these factors.C/FpH values for the sulphides. i. Modelled field and column rates were predicted according to the proposed equation (9). respectively (O pyrite. application of the procedure to further on parameters and mechanistic or empirical site data may indicate whether or not the proce- relations that are valid for the specific biologi. the first relatively successful attempt the mobile contamination that leaves the site . as well as implicitly helps to remove the discrepancy when predicting account for potential diffusional resistance to rates of weathering at field sites from laboratory weathering in large particles.and column-scale weathering rates plotted against observed weathering rates (R C/F in Table 2). Experimental deter. represents a relatively general set of factors that ering may predominate.e. If variation between sites is small and the flow is independent of the nature of the procedure is insensitive to these factors. it is important to recognize that the accumulated. Additional detail and numerical results not listed here are available as supporting information in Malmstrom et al (2000). which scales up the batch rate coefficients. A -logBC/FPS. The inserts show the logarithmic values of the scaling factors for the field (insert A) and the columns (insert B) according to: A . the This scaling procedure accounts for retention quantification of the particle size effect is of dissolved weathering products in immobile empirical and site specific. A -log BC/FT. D denotes overlap of the — logbeta. to our the risk assessment framework. Closing comments This contamination must be considered within The scaling procedure described above is. and if these cally mediated reactions at the site. to account specifically for the physical and bially mediated sulphide mineral weathering geochemical factors giving rise to the commonly (Stromberg & Banwart 1994). biotic and abiotic weathering processes. the actual quantification of the distribution. A plagioclase).C/FT due to negative — log beta. However.

BANWART.. The work originated from research & Postma 1993). Hydro- cesses.A.L. A. The project was led by interpretation of solute mass flows. Release Site. In: KOVAR. JAMBOR. Rot- ally create a significant error with respect to terdam. collaboration on the Aitik mine site by S. 55. such as surface water catchments. These include appli. that we consider at present to be more appropriate for Tier 2 assessment. If References neglecting reactions such as sorption and APPELO. International Association of Hydrological Sciences. with the discharge. and thus the relative retention of dissolved 965-978. This applies as well to any weathering products is similar between sites.A. ASTM Designation: D5744-96-Stan- general conceptual models generally allow dard Test Method for Accelerated Weathering of broader application to a wider range of site. Efterbehandling av sandmagasin allow greater use of tabulated fundamental data och grabergsupplag i Aitik. which include mineral solubi. and Solid Materials Using a Modified Humidity Cell. Croxford is conceptual site model. G. For chemical kinetic pro. Golder Associates reactive transport. K. now at Boliden Minerals Inc. fundamental data for specific min. respect to protons and activation energies. S. B. Publication.. The greatest question at field scale. Conshohocken. ASTM. reaction orders with 74. American Society for Testing and As stated in the opening section of this paper. procedure can be tested by prediction of weath. ering rates at other mine sites and for other BLOWES. M. D. DESTOUNI. BANWART ET AL. from M. Wallingford. range of reactions. The Quality: Remediation and Protection. additional site data. such as sulphide and silicate mineral chemical modelling for preliminary assessment of weathering. and contamination that is sorbed or is accumulated as whether neglecting retention of sorbed weath- secondary mineral phases and thus immobilized ering products and secondary minerals will within the site. In order to provide supported by a Hussein-Farmey Endowed PhD quantitative results. CHERRY. Journal of Geochemial Exploration. However. Hydrogeologiska when parameterizing mathematical models of fdrutsdttningarfor atgdrdsplan. (ed) Groundwater present is the generality of the approach.J. Their greatest at the Swedish Nuclear Energy Inspectorate) and advantage is the ability to explore a wider N. T. Groundwater and Pollution. 2001. C.. A. E. J. West the same allocation of resources. & MALMSTROM.). Stockholm. West Conshohocken. Special Issue on Mine Water erals and weathering mechanisms are available Geochemistry.156 S. PA.A. Destouni (KTH) with significant research input such codes do require a higher level of geochem. 250. 1996. BANWART. create sufficient errors at other sites to ensure The hydrochemical approach advocated here that they must be included in a general scaling is the simplest possible. 1995.. ASTM Designation: E1739-95 . C. they generally require Studentship at the University of Sheffield. AB. 1991. Report 927-1801. Ban wart during his appointment at the Royal Institute of Tech- lity equilibria as additional constraints for nology (KTH). mine water pollution. Evans is supported by the NERC Environmental secondary mineral formation. J. S. The scaling procedure described above in. now ical understanding of the system.. Geochimica Cosmochimica Acta. Stromberg (KTH. G. REARDON. then such ASTM.J. Balkema. with input varied quickly in order to test a range of conceptual site models. Key uncertainties importance of cemented layers in inactive sulfide include whether the fraction of immobile water mine tailings. The relatively good agreement Tier 1 assessment. including rate constants. (in Swedish). & MALMSTROM. & POSTMA.A. J. 1992. BYSTROM. 1993. & systems. In addition.Guide processes will need to be conceptualized and to Risk-based Corrective Action at Petroleum included in the scaling procedure. such as characterization of secondary minerals. secondary mineral formation is found to gener. when testing the Diagnostics Thematic Programme. Materials. D. application to Tier 1 assessment. dependence of mineral weathering rates between chemical codes such as PHREEQC (see Appelo laboratory and field. 1998. It is clear that more advanced modelling The authors wish to acknowledge the original work approaches can be applied. 73-97. S. PA. Geochemistry. HOLMEN. conceptual models that are based on fundamental AXELSSON. The formation and potential where suitable datasets exist. Malmstrom (KTH). including sorption and K.W. J. A major advantage of the between laboratory and field results for minerals approach is its transparency. M. and is appropriate for procedure.L. Assessing mine water pollution: from laboratory to cludes these factors. 307-311. allow greater depth of site interpretation for American Society for Testing and Materials. & JANS- descriptions of reaction and transport processes SON. Eriksson (KTH. . Results can be other than chalcopyrite suggests this is not the obtained easily with spreadsheet calculations case for the Aitik site. and publications cited here on resolving the scale cation of computational thermodynamic geo.

In: MELCHOIR. US Contaminated Land Centre. Journal of Environmental Quality. W. Comparative analysis of laboratory and field tracer STROMBERG. . Energy and Minerals Groundwater Quality. Applied Geochemistry. In: STUMM. & of Mines and Quarries. FREEMAN. L.S. PL. Sulihull. Cont MALMSTROM. ommended Methods for the Prediction of Metal YOUNGER. SCHULLER. B. Environment Agency and Remediation SOBEK. DC. GUPTA. \999b. D.L. in press. & BANWART. composition in waste rock dumps undergoing Sweden: scale dependent rate factors and pH pyritic oxidation. (eds) 1995.L.. Symposium Series. A.J.R.A. J. 416. Environmental WHITE. J. Pore gas kinetics of waste rock from the Aitik copper mine. 1990. 5.B. 59-89. Smithers. 1985. WHITE. 143-163. 1990. 143-152.. & DESTOUNI. OSHAY. S. B. Resolving the scaledepen. & PETERSON. Pollution. 1997. 1986.. Reclamation particle size..R. International Association of Division. & BANWART. 583-595. In: rates. STUMM. controls in large column experiments. John Wiley. British Columbia Ministry of nated groundwaters at abandoned mine sites. 1990. Geochemistry. Weathering Rates of Silicate Minerals. & SOUTHAM. Predicting temporal changes in for determination of potential acidity in pyritic total iron concentrations in groundwaters flow- overburden.. (eds) Environ- Kinetics. & GIFFIN. & BASSETT. STROMBERG. BANWART. BC. biology of sulfide mineral oxidation. Reviews in NORDSTROM.R. Kinetic tests for investigating preferential flow and modelling of geochemical processes at the Aitik transport in mining waste rock.L. A. 14. Aqueous Systems II. J. PL. G. S. 1997. DESTOUNI. dence of mineral weathering rates. S. & BANWART. & BANWART. 1-16. Remediation. W. Field and Laboratory Methods England and Wales. 1990. N. 1999a. New York. D. Washington.B.A. T.. 2001. 19. Journal of Contaminant Hydrology. Soil Science. Joural of Hydrol. M. Environmental Protection Agency Publication.K. 2002. 194. HOSSNER. G. A American Chemical Society. Environment Agency of SMITH.M. A comparison of laboratory and field weathering Mine Water Hydrology. M. 1375-1377. & RITCHIE. Time-scale Leaching and Acid Rock Drainage at Minesites in issues in the remediation of pervasively-contami- British Columbia. logy: interactions between microbes and minerals. ERIKSSON. R.A. PREDICTING WEATHERING RATES 157 BONE. Wallingford. W. 361-390. vol 31. STROMBERG.A. 44. J. S. Kluwer. & HEDIN. and Revegetation Research. 1997. 2002. S. 140. 35. (eds) Chemical Modeling of Reviews in Mineralogy.L. Weathering HARRIES. John Wiley. Applied ogy.. 2000. Dordrecht. Aquatic Chemistry. R. R. B. S. D. Geomicrobio. B.A.F. Hydrogen STROMBERG. mining waste rock site in northern Sweden. Geomicro- Mineralogy.) Aquatic Chemical ALLOWAY.J. B. & DIXON. DRAFT Guidelines and Rec. G. In: Employment and Investment. 39. A. (3rd edn). PRICE. A. Hydrological Sciences. Chemical Science & Technology. & MORGAN. A. Experimental peroxide oxidation: an improved method for study of acidity consuming processes in mining rapidly assessing acid-generation potential of waste rock: Some influences of mineralogy and sediments and sedimentary rocks.E. EPA-600/2-78-054. R.F. B. 2000. BANWART. & BRANTLEY. 34. FINKELMAN. mation.A. J. 1994. S.L. & TREVORS. J. SCHNOOR. mental Pollution Series. Kinetics of chemical weathering: YOUNGER. National Groundwater & Applicable to Overburdens and Mine soils. J. modified hydrogen peroxide oxidation method YOUNGER. PL. ing from abandoned deep mines: a first approxi- 778-782. 1978. 521-534. 47-69.I. & Hydrol. New York. (ed. 9. W. M.T.

.

The hydrogeochemical dynamics of mine pit lakes R. 159-185. However. MPLs tend Murray 1997. Davis & Eary 1996). 1).S. processes (Bird et al. direct precipitation most natural lakes. method for extraction in the mining industry in the predictions tend to be hypothetical.00 © The Geological Society of London 2002. The geochemical preserved through the development of a chemo. 1997. Special Publications. Further. P. which allows the economic utilization assessment of wall-rock behaviour (Pillard et al. J. 159 Geological Society. The prediction of future pit lake water quality within a MPL is. Windsor Court. used the pit lake on a seasonal migration. Consequently. 1996. In many cases the requirement for predictions The closure of open-pit mining operations of water quality is made during the design and requires careful planning to avoid environmental From: YOUNGER. This 1995. then on closure and abandonment of these Methodologies for the prediction of pit lake facilities. Consequently. sulphate and depressed pH. it may present a risk to the environment through the development of poor quality water with elevated concentrations of metals. biological and physical processes. and show changes related to sea. essential in considering environmental impact on a closed or abandoned mine facility. the pit fills with water producing a chemistry involve construction of a series of mine pit lake (MPL). whereas. based on recent years due to the advancement in mineral numerical predictive modelling and laboratory processing. therefore. and limnological and geochemical surface area ratio is usually greater than 1 for a processes operating within the MPL itself. The length of time needed linked empirical and physical models to rep- for the lake to reach full depth will depend on resent the major controlling physio-chemical several factors including pit dimensions. wall-rock-groundwater interaction. By comparison. BOWELL SRK Consulting. of near-surface low-grade high tonnage ores. N. and will involve chemical. London. metalloids. recharge is generally through level of exposure. Open-pit mining has become a common place permitting stage. 198. & ROBINS. desorption MPLs can and do exhibit thermal and chemical or adsorption of metals and metalloids with stratification. 2). . groundwater regime. Another major difference is that. These processes to show a greater depth to surface area ratio than include groundwater inflow. in some cases. pit lakes are smaller than water. 0305-8719/02/$15. This body of water remains as a permanent feature on the mine site and as such becomes a closure issue with respect to water quality and potential to degrade groundwater. for MPL can lead to organisms having a harmful natural lakes. Kirk et al. Localized and regional-scale processes affect these in turn. Cardiff CF10 3BX. For example. evap- natural closed lake basins and thus the depth to oration. fall-evaporation budget. prior to mining. in order to predict pit lake water quality it is essential to understand the hydrogeological.L. 342 migratory geese were exposed flow is generally more important (Atkins et al to a lethal concentration of metals when they 1997). Eary 1998). with implications for water and pit wall-precipitation contact or run-off circulation. 1-3 Windsor Place. available for the model (Kempton et al. within the surface water. the greater depth to as well as organic processes such as methylation surface ratio limits water circulation and thus a or the activity of sulphate-reducing or sulphide- permanent stratified or anaerobic layer remains oxidizing bacteria (Fig. 1994. Kempton et al. It is particularly true for gold and copper operations follows that. The MPL (Table 1). geochemical and limnological processes that influence water quality. Miller et al 1996. The controls on a MPL will vary over time. (eds) 2002. for MPLs. rain. sonal turnover. groundwater inflow and Havis & Worthington 1997. the predictions can in many parts of the world (Peters 1987). 1997). be in error due to limitations in the input data If mining has occurred below the water table. groundwater Berkeley Pit. UK Abstract: On cessation of mining open pits or opencast workings that extend below the water table are likely to fill with water and thus develop a mine pit lake (MPL). precipitation and dissolution. mineral surfaces. enrichment of potentially toxic elements in a line (Fig. latter includes oxidation-reduction. Mine Water Hydrogeology and Geochemistry. Generally.

33 5. the what is the relationship of the MPL to the assessment of water chemistry is a critical aspect environment.510 1367 22 319 Pyramid Lake 438 102 4 294 Mine pit lakes Berkeley Pit 0.32 259 1.084 17 4.94 Magcobar 0. 2.485 455 152 714 Lake Baikal 30. the fundamental consideration is resources in a mining area. This paper presents an invited review of the current status of knowledge regarding the major hydrogeochem- ical processes that operate within a MPL. . It should be noted that hydrogeological and hydrological processes are also important but are outside the scope of this paper and will not be discussed in detail here.160 R.158 70 2. the formation of a MPL can affect In considering the interaction of a MPL with its both the availability and the quality of water environment. Conceptual model impact. 1.113 21. 1999 with additional data) Surface area (km2) Depth (m) Surface area/depth ratio Natural lakes Lake Superior 82. Fig.55 Crone Bane 0.38 Ruth Pit 0.222 40 5. Schematic representation of the geochemical processes that influence water quality within a mine pit lake. ture used in referring to the zones within a stratified pit lake (after Drever 1988).23 Summer Camp Pit 0.350 395 208 481 Lake Victoria 69. Thus. Shevenell & Pasternak (2000) and papers else- Fig. Schematic representation showing nomencla. The reader is referred to Fetter (1994). Comparison of hydromorphic properties of pit lakes with those of Natural lakes (After Tempel et al. where in this volume. If the MPL acts as a terminal sump. Havis & Worthington (1997). J. BOWELL Table 1. Consequently.26 of a open-pit closure programme.

45-um fraction includes all metals mobilized 1997). 5) or anions (Fig. and groundwater shows a by high concentration of dissolved metals as similar chemistry to MPLs (Davis et al 1998). Miller et al 1996. The chemistry of a MPL may change with time and reflects a dynamic rather than static equilibrium. high metalloid ment along with water draining old underground concentrations can occur (Fig. impacted the immediate groundwater environ. but these represent the more Getchell. can exceed pH 7. the lowest hydrological point within a basin. Nevada. outflow other than evaporation (Fig. In the Robinson North Pit on the Getchell Mine in Nevada. For example. has recently been published by Eary (1999).45 um) rather than total. such as at Magcobar mine in then it acts as an evaporation pan with no net Ireland (Table 2). High sulphide (particularly pyritic) wall- rocks. physico-chemical status in controlling . 1997. tend to produce poor water quality (Table 2). Consequently the chemistry of MPL can vary widely and this can be related to three major groups of controls: • limnological processes . such as those exposed in the Parys Mountain and Berkeley pits. such as at Magcobar in Ireland. In some cases. • geochemical constraints . In many cases MPL chemistry can be some pit lakes. the pH of the pit lake common ones. and hydrochemistry. Davis et al. hydrated cations and chelates (Fig. The Berkeley Pit at Butte. and oxyanion concentrations are quoted from through' can occur within the MPL and it can filtered samples (at 0. These are shown in Table 3 and are supplemented assumed to be acidic. Yerrington. Robins et al 0. relevant to every site. the chemistry of Fig. Ely. such as the basal sediments of (Table 2). where water chemistry has changed from acidic to circumneutral (see the case study below). Oxidized wall- rocks that contain appreciable carbonate tend to produce better quality pit lakes due to the abundance of reactive buffering in the host rocks. In high pH environments. The major mineralogical controls on pit lake chemistry. at the Summer Camp Pit site in Nevada. Certainly in these waters evapo- oxidized sulphide-bearing host rocks both in concentration is an important mechanism for underground workings and open pits that distinct enrichment of As and also for Se (Eary 1998). due to local variations in such as those as Copper Flat.e. 3. Not all of these controls will be circumneutral and have low metal chemistry. and has form. metal concentrations will include paniculate 1998). It should be noted that in this paper the metal However. This is because total environment (Robins et al. 6). such as in the workings (Robins et al 1997). if the MPL is not a sump then 'flow. This pit lake is acidic and has high metal as free ions. such as those pit lakes with phases that are important in reducing developed at Parys Mountain or Berkeley environments. erroneously.this is related to the minerals present. By hydrogeochemical groundwater blocks can be contrast. where Mining District. The Ashenberg 1989. Hydrogeological scenarios for a mine pit lake reactive water and the environment's interaction with the the groundwater environment. 4). chelated species and in a colloidal and sulphate concentrations (Table 2). 3). Montana has metals as well and thus the influence of chemical been one of the most studied MPLs (Davis & release processes could be overestimated. HYDROGEOCHEMICAL DYNAMICS OF MINE PIT LAKES 161 i. low pH environments are characterized mapped in the district. interact with the surrounding groundwater unless otherwise stated.this relates to the lake's physical processes and how they will influence geochemical reactions. based on predictive thermodyn- Mine pit lake chemistry amic modelling for a range of Nevada pit lakes. groundwater has historically As levels can reach 10 mg 1–1 (Miller been sufficiently influenced by reaction with et al 1996). The chemistry of a MPL is often.

1 ) 0 0 512 0 0 0 .05 < 0.134 Pb (mgl ') < 0. (1996) (1989) pH 2. North Pit. Magcobar. - S0 4 (mgl .8 65.005 < 0.3 435 280 32.002 Se (mgl ') 0. (1996) data data data Ashenberg (1996) (1996) et al. 0.16 0.007 < 0. Bowell Unpublished Unpublished Unpublished Davis & Miller et al.6 0.2 0.006 245 156 37.2 Ca(mgl-1) 21. Miller et al.1 ) < 0. mg/L 4.02 0.001 < 0.8 3.005 0. Corta Atalaya.2 Ba(mg1 .67 8.005 22.4 0. (mg1 .005 Fe (mgl ') 1300 0.001 8.9 7.1 < 0.2 0.6 586 161 379 462 558 530 45.005 0. Ruth Pit.73 344 95 68.1 ) 5200 2530 563 14300 5740 2840 1570 90.003 - Zn (mgl ') 54 29. 0.002 . Cortez Pit.004 0.13 0. < 0.6 2.32 0.005 0.6 23.11 .1 ) 6000 3750 1210 15000 .77 7 1. Berkeley Pit.032 4.77 < 0. Mountain Ireland Ireland Rio Tinto.02 < 0. Representative analysis of mine pit lakes Parameter Parys Crone Bane.07 TDS.1 4700 386 6.9 .005 < 0. Miller et al.071 < 0.4 0.08 5. Nevada (1982) Nevada Ref. 0.7 .38 0.001 < 0.6 2400 . Cortez.008 . - As (mgl ') 0.3 5.Table 2.06 A l ( m g l 1) 108 11 0.5 6.004 Mn.005 0. 23. Spain Montana Nevada (1993) Getchell Mine.038 Cu (mgl ') 69.001 . 4150 2420 432 Alkalinity ( m g l .

HYDROGEOCHEMICAL DYNAMICS OF MINE PIT LAKES 163 Fig. Flicklin diagram assessing variation in divalent metal cations with pH and ore deposit type (Flicklin et al 1992). Variation in arsenic within mine pit lakes as a function of ore deposit type (using the classification of Du Bray 1995) and pH. Fig. Data from published and unpublished sources. 6. Younger diagram (Younger 1995) for various pit lake waters. 4. . Data from published and unpublished sources. Data from published and unpublished sources. Fig. 5.

5) Aluminium Alunite. Geochemical controls on major and important trace components in a mine pit lake system (based on a scheme formulated by Eary 1999).5) Circumneutral (pH 4. ' As a trace element.and Fe-sulphides* Barium Barite Barite Barite/witherite L E W O B J.5-7.Table 3. basaluminute Gibbsite Gibbsite Alkalinity Not applicable Calcite Calcite Arsenic Scorodite/adsorption onto HFO Adsorption onto HFO/Fe-sulphides* None identified/As. nickeline Selenium Gypsum Adsorption onto HFO None identified Sulphate Gypsum Gypsum Gypsum Zinc Zinc melanterite Hydrozincite/smithsonite/ZnSiOyZn-sulphides'* Zincite/Zn 2 SiO 4 /Zn-sulphides* * In reduced environments. Prediction of mineral phase control based on thermodynamic calculations as well as mineralogy of pit lake sediments Component Acidic (pH < 4.5) Alkaline (pH > 7. . + Cadmium None identified Adsorption onto HFO/Zn-sulphides* Otavite/Zn-sulphides* Calcium Gypsum Gypsum Gypsum/calcite Copper Chalcanthite/cupromelanterite Brochantite/covellite/Fe-sulphides* Malachite/brochantite/covellite* Fluoride Fluorite Fluorite Fluorite Iron Melanterite/halotrichite/Fe-sulphides* HFO/ferrihydrite/Fe-sulphides* HFO/ferrihydrite/Fe-sulphides* Lead Anglesite/chloropyromorphite Anglesite/chloropyromorphite/galena* Cerrusite/chloropyromorphite Manganese Manganite/birnessite MnHPO4 Rhodochrosite Nickel Morenosite/Fe-minerals+ Morenosite/Fe-minerals' Gaspeite/ni-oxides.

ground and ground or permanent. lake may overturn in late summer or fall even At some point in the late autumn or early before the surface water temperature drops below winter. temperature.this is related to the The water behaviour during winter depends on characteristics of individual mineral depo. The density of water is a function of thermal conductivity. Overturn through the water column TDS. water density increases with increas. the temperature of had formed in the winter it gradually melts and maximum density of water also decreases. As the air temperature and solar density stratification. decreases rapidly with depth. In spring. because wind-induced mixing concentration at the lake surface can increase the in mid-latitude pit lakes seldom reaches beyond TDS and thus the density of the surface water. 1). In dry temperate climates that experience oxygen. stratification can occur if a lake has a significant Eventually. the surface water temperature drops the temperature of the underlying water. the water temperature increases from the surface downward. the entire lake may remain well densest at a temperature of about 4°C. This heating causes the density of the surface layer to decrease. the underlying water. if the lake weeks. as long below that of the underlying water and the as the density of the surface water exceeds that of surface water density is now greater than that of the underlying water. water. • geological controls . depending on the temperature structure of in the TDS of various source waters to the lake the lower layer and the amount of cooling at the and/or to processes in the lake that increase the surface. The surface water sinks Long-term (multi-year) or permanent density until it reaches the level of its new density. When the surface water temperature reaches 4°C. In such cases. radiation increase. potentially to the bottom of the lake. Fresh water is vicinity of 4°C. whether ice forms. the surface water is heated (Fig. separates these the addition could induce stratification near the layers (Fig. If ice ing TDS. the water below sits and will influence the extent to which the ice may gradually lose its oxygen by the same geochemical reactions will occur. even when no temperature differ- surface water is mixed by wind or surface inflow ences exist due to the lower density of the snow- and is typically in equilibrium with atmospheric melt. At a given mixed and oxygenated throughout the winter. If no ice had formed and tion in mid-latitude lakes of uniform dilute the lake had continuously overturned in the chemistry (uniform low TDS) is due to increases winter. For lower temperature. The epilimnion is typically less than positive net evaporation in the summer. processes that occur in the lower layer in summer. As its TDS increases. Overturn causes mixing that replenishes contains enough organic matter to deplete oxygen . If ice forms. warmer temperatures in the spring would in ambient temperature in the spring and repeat the cycle of thermally induced seasonal summer. the denser surface water. Pit lakes in temperate climatic zones can develop Which process occurs depends on the relative vertical density stratification that may be seasonal temperatures of the air. oxygen may This density increase can potentially offset the gradually become depleted by oxidation of dead density decrease of the surface water due to algae and lake fauna that fall into the hypolimn. The lake then has a Mine pit lake surface layer ('epilimnion') of uniform lower The seasonal cycle just discussed assumes that density and higher temperature. HYDROGEOCHEMICAL DYNAMICS OF MINE PIT LAKES 165 the extent to which release and attenuation the oxygen throughout the depth of overturn. evapo- about 10m deep. thermal heating in the summer. example. the pit lake has essentially no TDS and its density lying layer ('hypolimnion') of higher density and is unaffected by precipitation or evaporation. the entire water column may over. this depth. Below the ice. For example. the air temperature increases. Below the thermocline. the Natural lake processes lake may again overturn and reoxygenate due to Thermally induced seasonal density stratifica. 8). as well as from the bottom up. This in turn affects the density of the could take less than 1 day or occur over several deeper water (Fig. the ion from the surface. vertical variation in TDS due to large differences turn. If no ice forms and the air both its temperature and its salinity or total temperature keeps the surface water in the dissolved solids (TDS) content. if spring snowmelt and run-off have a cline' or metalimnion. the water may become temperature stratified with 4°C water at the Limnological processes bottom or it could remain well mixed due to heating from the sides and bottom of the pit. Above the thermocline. A zone called the 'thermo. mechanisms can proceed. the lake surface. and an under. in which the temperature significantly lower TDS than the lake water TDS. 7).

in which the TDS ferric hydroxide precipitates falling into it from increases rapidly with depth. then during the summer deep (well below the depth of summer wind ferric hydroxide that precipitates at the surface mixing).166 R. BOWELL Fig. J. A zone remains so and will continuously dissolve any called the 'chemocline'. If the bottom layer is sufficiently of the hypolimnion and strengthens the density . Thermally induced stratification within a lake (from Parshley el al. upper layer of lower TDS water overlying a layer once the hypolimnion becomes anoxic.). a chemically stratified lake has an metalimnion. Thus. separates the two above. it of higher TDS water ('hypolimnion'). raising the TDS content and above the denser hypolimnion. in the hypolimnion. autumn-winter usually occurs throughout the Typically. Overturn in the the density of the bottom water. This process further increases the TDS layers (Table 4). a seasonal middle layer ('metalimnion') will sink. 7. become reduced and dissolve in the may form below the summer thermocline and basal anoxic water. but not the hypolimnion. in prep.

Atkins et al 1997. Geochemical controls on MPL chemistry perpetuating the stratification. • Attenuation processes . Parshley elements into the pit lake. processes can below).). 8. . to accumulation of elements in the MPL. similar trends indicating that turnover occurs in • Concentration processes . Chemical and thermally induced stratification (from Parshley et al. 9). be viewed in terms of: Few studies reporting MPL site-specific limnological data have been published to date • Release processes . will lead to natural attenuation In considering the geochemical processes that of metals and metalloids as well as sulphur (see influence MPL chemistry (Fig. HYDROGEOCHEMICAL DYNAMICS OF MINE PIT LAKES 167 Fig. submitted.those that release (Gannon et al 1996. gradient between it and the overlying layer. submitted). in the hypolimnion. Sulphidization. These studies deal with pit elements from the hydrosphere either as lakes in Nevada in the western USA and display precipitation or through adsorption.those that lead MPLs similar to natural lakes.those that remove et al.

005 1 7.016 0.8 1.01 15.078 0.06 1.9 Dissolved O2 1.15 4.002 0.006 Pb 0.2 0.01 16 0.03 4 As (III) 0.6 0.01 0.005 < 0.1 158 Fe (total) <0.53 0.013 0.004 < 0.018 0.3 As (total) < 0.5 0.003 Zn 5.7 6.96 Cu 0. .6 7.025 16 Cd 0.4 Alkalinity 451 753 74 0 154 85 Chloride 46 51 13 59 Sulphate 370 2200 22 420 Na 26 30 33 36 25 34 Mg 52 82 16 46 7.2 3.4 15.001 0.7 6.19 Fe (II) 0.54 7.018 0.8 8 0.67 8.2 0.1 0.2 160 Ni 0.003 < 0.1 unless otherwise stated.21 0.01 3.002 0.8 16.1 8.14 414 0.2 26 0.Table 4.5 Eh (mV) 278 —108 256 160 TDS 555 885 714 3490 245 922 Temperature °(C) 7.007 0.01 6.9 2.058 3.001 * All concentrations in mg 1 .9 3.65 416 1.6 0. (m) 0 50 0 23 0 20 pH 6. Examples of stratified pit lakes Date Magobar Magobar SCP SCP SCP SCP February 1998 February 1998 25 February 1998 25 February 1998 13 June 1998 13 June 1998 Depth.21 0.003 < 0.007 0.5 19 Ca 142 242 140 400 30 160 Mn 6.

where X is the adsorption site C. Relationship between release-attenuation-concentration processes.OH (goethite)+(n+l)H2O Year(s) . A general- This involves mineral-water reactions such as ized set of relative half-time reaction rates are sulphide oxidation. 10-6 s Solute.water reactions: CO2+H2O = H2CO3 (hydration) C. s-h Oxidation-reduction : Fe2++0. Principally these involve sorption onto The primary leaching processes for metals in- colloids or evapoconcentration of salt clude sulphide oxidation and associated mineral components.n H2O = a-FeO. HYDROGEOCHEMICAL DYNAMICS OF MINE PIT LAKES 169 Fig. The rates at which such reactions occur are important as they can occur over a time scale of seconds to Release processes millions of years (Langmuir 1997).25O2+2. 10-6 s Aln+m(OH)3n+2m+m H2O = (n+m)Al(OH)3(s)+mH1 (multivalent ion hydrolysis) h-year Adsorption/desorption reactions: X(AsO34-)+PO34.= X(PO34-)+AsO34–. sulphate dissolution or shown in Table 5.l s Cu2++2H2O = Cu(OH)2+2H+ (hydrolysis/complexation) C. General reaction rates important in mine pit lake systems (based on a compilation by Langmuir 1997) Reaction type Reaction time frame Solute reactions: H2CO3 =H++HCO3 C. The wide range of precipitation rates reflect Table 5. 9.0. hydrolysis mineral buffering of water chemistry. Within reaction rates occur at relatively fast rates many hard rock pits the most important reactions while mineral dissolution can take much longer. 10 -10 s Fe3++3H2O = Fe(OH)3+3H+ (hydrolysis/complexation) C. are those involving sulphide-bearing rocks. both of which increase the total dis- solved solid load in the resulting water. As can be observed. buffering.5H2O = Fe(OH)3+2H+ min-h Mineral–water equilibria: Ca2++HCO3– = CaCO3+H+ week-year Isotopic exchange: 34 SO24+H32S– = H34S–+32SO24– Year(s) Mineral recrystallization: Fe(OH)3.

Alternatively. of the total acidity and store some acidity in • the concentration and presence of acid. portion of the available hydrogen is released. they can release only a portion generating phases. muir 1997). most noticeably in the case of MS2 (where M is a metal cation. these This is a first-order reaction with a rate given minerals are important as both sinks and sources as l X 1 0 – 4 s . • the concentration and presence of acid. With oxygen as the oxidizing precipitation and rapid release on exposure to agent the rate can be considerably faster by moisture (Nordstrom 1982. Cravotta 1994). interface. These minerals are highly soluble so supplied and reduced at the cathode-solution can represent an instantaneous source of acidic. Essentially. The oxidation of pyrite by ferric iron. Bowell et al 2000). for example the dissolution of anodic. BOWELL the variation between precipitation of amorphous consume H+ on oxidation. possibly. J. Mo) and/or with iron-bearing sulphides (Thornber 1992). it should be noted that not all eral buffering of H+ in drainage. sulphate and. Fe3+. For each mole of pyrite oxidized. To reflect this the solids that can occur in a matter of minutes to oxidation of chalcocite can be written as: formation of crystals that occurs on a geological time scale. When sulphides are exposed to oxygen and water in the presence of a catalyst they undergo oxidation. Anodic reactions can occur deep within cracks. alkali-producing reaction from the hydrolysis. of acidity. Dissolution or oxidation of pyrite initially or saturation to form secondary minerals produces Fe2+ that is almost instantly oxidized to (Thornber 1975. which are stable only in oxidiz- consuming phases. which is either precipitated as an oxyhydr- Alpers et al 1994. the determination of the resulting water chem. and increased acidity. Consequently. 1992. frequently accompanied by the precipitation of ber 1992. then the more extensive the conducting area. On weathering. acid-producing reaction will melanterite: have a major control on the mineralogy of the resulting assemblage. metal ions on Moses et al 1987). The greater the distance between cathode and anode. metal sulphate-rich water upon dissolution and consuming. only a conductor-metallic range (Thornber 1983.1 (McKibben & Barnes 1986. oxidizing. Miller et al 1996. Jennings et al 2000). Fillipek et al 1988. sulphides can release all acid istry in most hard rock scenarios will depend on: potential precipitating secondary minerals (Wil- liams 1990. eg Cu. The rate of sulphide oxidation can be rest is stored as partly oxidized metal sulphate controlled by the rate at which oxygen is minerals. Nickel & Daniels 1985. sulphate and possibly protons. The mechanisms of sulphide oxidation involve the transfer of electrons because most sulphide minerals are electrical conductors in the semi. can be given as: The ability of these minerals to react with water will depend on solubility. Nordstrom oxide or reduced by pyrite generating more Fe2+ & Alpers 1999a. Oxidation of ferrous iron and hydrolysis of fissures and along grain boundaries where ferric iron at pH > 2 provide the additional solutions can penetrate without the necessity source of acidity through the reactions: for dissolved oxygen (Thornber 1975. 1992).170 R. Nordstrom & Alpers . Acid-neutralization reactions result from min- However. sulphides secondary minerals (Kwong & Ferguson 1997. orders of magnitude (Karamenko 1969. Hence. for example the for- Sulphide-water reactions can produce acid mation of melanterite: pH water. The reaction leads to release of metals/metalloids. This buffering is sulphides on oxidation generate acidity (Thorn. secondary salts. this leads to greater potential for sulphide oxidation. Indeed. of the type M2S such as chalcocite actually Lawrence & Wang 1997. Lang. ing acidic pH environments. The 1992). b). The separation of the cathodic oxygen. Nordstrom & Alpers 1999a.

Sverdrup HFO is positively charged in the Helmholtz layer (1990) divided the most common minerals into (Deng & Stumm 1994). 1994. Carbonate minerals (especially et al 1994. phosphates. hydrous ferric oxides precipitate that Silicate weathering as a proton sink has been may ultimately crystallize to form goethite or a demonstrated in previous studies (Sverdrup similar ferric hydroxide (Bigham 1994). Scott 1987. 11). of the precipitated mineral assemblage is pH. and particularly the level of ferrous iron in these minerals is an additional As and heavy metals. negative. 10). due to the excess of their sluggish reaction rates. These reactions can reduce acid gener. Moss & Edmunds increases. However. sulphates or solved concentrations can increase in the pit lake hydroxides. Deng & Stumm of acid and precipitation of metal hydroxides. Bowell et al 1996). for example in the case authigenic sulphides not observed in bedrock geo- of copper (Fig. ankerite the (attracting cations). In low pH environments six groups according to pH dependency of their these HFO particles are usually colloidal sized dissolution rate (Table 6). Under acidic conditions. above 3. noticeably the reactive sulphides. Kwong & Ferguson 1997). From soil acidification studies. as they are not strongly adsorbed. like solubility.g. to a lesser extent. goethite constraints and partly by kinetic limitations and jarosite amongst others. Fuge (Morse 1983). as water pH increases minerals. Many oxide The order of carbonate-neutralizing capacity is surfaces change from being positive at low pH calcite > dolomite > ankerite > siderite. be noted that in many acidic as the only geological source of neutralization environments. however. or HFO (this may also is partly controlled by equilibrium mass action include minerals like schwertmannite. This order of reactivity hydrous ferric oxide. as the surface of the sidered. Bigham 1994. dolomite and element binding at the mineral-solution inter- magnesite) readily dissolve and provide bicar. has been shown to be source of acidity due to the strong hydrolysis of influenced by adsorption onto precipitated ferrous iron in solution. Mineralogical commonly the main precipitates. In the groups. The point at which this occurs is termed the point of zero charge (Stumm 1990). silicate and pH. destabilization as aqueous species. At low buffering capacity of mine wastes. while at higher analysis of pit lake sediments at Summer Camp pH other salts such as carbonates and hydroxides Pit. is pH-dependent (Sigg & bonate alkalinity that results in neutralization Stumm 1980. composed of silicates and hydroxide-oxide As described above. Stumm 1992. reason for the limited neutralizing capacity is that Pit lake chemistry. including mackinawite ([Fe.Co. face and. Nevada (see case study below) has identified become more abundant. iron hydroxides and clays. carbonates dominate HFO can lead to high As concentrations in only limestone. oxyhydroxides and sulphates are sulphidization in the hypolimnion. Even for minerals hydroxyl molecules. to saturation with respect to some species. carbonates. these com. ferric hydroxide solubility decreases 1992. Bhatti et al 1994. HYDROGEOCHEMICAL DYNAMICS OF MINE PIT LAKES 171 \999a). As the pH increases and colloid (Table 6). As pH 1990. To assess the with a minimum being around pH 6-7. precipitated HFO tends to scavenge nega- hydroxide minerals therefore must also be con. solution. their dis- arsenates. Some solutes can be attenuated through ation by forming an inhibitory surface coating on adsorption onto mineral surfaces. Ni. This is the process of carbonate minerals (e. Dzombadt & Morel 1990).Zn]9S8) . tively charged oxyanions. as well as competition from complexing whilst the majority of geological materials are ions mobilizing As-oxyanions. It calcite) have often erroneously been thought of should. they will not be practical neutralizing case of goethite this occurs at a pH of between 6 materials unless they occur in excess of approxi. An important control on the diversity from continued release and evaporation. The accumulation of solutes in solution will lead In circumneutral-alkaline oxic environments. Attenuation of metals also occurs through At low pH. Hiemstra & van Riemsdijk mately 10% (Sverdrup 1990). 11). flushing or dissolution of these potential (NP). logy. From the relative and have a high reactivity proportional to surface weathering rates of the mineral groups shown area (Fig. minerals in the poor-negligible neu. in response to either saturation or and HnSeO4—(2—n) respectively (Bowell 1994). 1996). dolomite and marble rock types. As pH increases the surface of the HFO particles attracts Attenuation processes metallic cations and releases oxyanions (Fig. and becomes. the surface pH of the in the intermediate and fast mineral weathering particles changes. particles aggregate as Fe-OH bonds become tralizing categories are unlikely to react due to longer and more rigid. These molecules do form sparingly soluble solids pounds precipitate as secondary minerals such as and. and 9 (Parfait 1978. In the (thus attracting anions) to negative at high pH case of siderite and. As and Se form species such as HnAsO4—(3–n) Consequently. calcite.

. serpentine. zoiste. nepheline.9 0.7-5. actinolite. Intermediate weathering Epidote.3-3. 4.004 * Buffering pH range evaluated by crushing 5 g of pure mineral and mixing with 5 ml of distilled water and left to react for 30 min. 2. tremolite. talc.1 0. 5. NP range* Relative reactivity* 1.1 g of calcite. ferrihydrite and muscovite 2.2 7.6 0. oligoclase. aragonite. Whereas 10 g of hornblende was required to buffer HC1 acid to a similar pH to only 3.3 1.Table 6.01 0. see below and based on 100% mono-mineral sample. anthophyllite. magnesite. aragonite.4–5. Fast weathering Anorthite. rutile and zircon 3. vermiculite.2-4. + NPrange assessed as equivalent buffering potential of 10 g of pure mineral to calcite and titrated with hydrochloric acid.8-14. 6-11. olivine. enstatite.1 0.8 1.02 manganite. montmorillonite. hornblende. Slow weathering Albite. 10 g of portlandite (Ca(OH)2) was found to have the equivalent capacity to neutralize HCl acid as 14. hyperthene.0 portlandite and brucite 2.8-0. glaucophane. siderite and wollastonite 3.5-11 2.6. augite. gibbsite and kaolinite 5. Grouping of minerals according to their neutralization potential (Sverdrup 1990) Group name Typical minerals Buffering pH range* Approx. garnet. kutnahorite diopside. The pH of the distilled water was 3.01 6.8 g of calcite. So. jadeite. for example. hematite.4 hedenbergite. + Calculated from Sverdrup's (1990) equation.2-0.8 0.6 leucite. labradorite. 4.4. chrysotile. goethite.2 0. Inert Quartz. dolomite. biotite. Dissolving Calcite. chlorite.5 <0. Very slow weathering K-feldspar.5-2.8-7. spodumene.

When Ca concentration in minimal then calcite precipi- tation will reduce to such a small amount that carbonate-bicarbonate molecules will concen- trate in solution and exsolve CO2 with further evaporation. will be the next control on pit lake chemistry. which in the case of most MPL is that of gypsum. This concept of 'chemical highly alkaline conditions (pH > 9). This will increase water pH and the resulting highly saline alkaline fluid will have a chemistry of Na-K-HCO3-CO3 Cl SO4. 12). In this divides' has become the basis of understanding environment oxyanions are not adsorbed and can the influence of evaporation on natural lake chem. Eary (1998) predicted that equili- precipitation. This can be expressed as: Consequently. HYDROGEOCHEMICAL DYNAMICS OF MINE PIT LAKES 173 calcite is precipitated. by a and greigite (Fe2+Fe32+S4). then evapoconcentration increases until Fig. By contrast. For most MPL In dry environments. evolves over time to become circumneutral Ca- neutral to alkaline. Drever 1988). Similar processes have reaction such as: been observed in deep-sea anoxic sediments (Morse 1994) and natural wetlands (Kwong & Van Stempvoort 1994. dominated by carbonate alkalinity and can the latter case will be the more common. Simplified Eh-pH diagram for the system Cu-S-Cl-O-H at 298K showing the dominance all available carbonate is removed by calcite fields of copper minerals. calcite will precipitate (Eary 1998). this is the first chemical divide observed in the development of lake chemistry at equilibrium with atmospheric CO2 (Drever 1988). the concentrations of Ca and bicarbonate (and thus pH) will be controlled by calcite precipitation and evapoconcentration. Nevada (Drever 1988). the latter can exert an important brium chemistry would be of the type circum- control on water chemistry. 2mca > malkalinity then pH is buffered to be If a MPL is a terminal sink then evapoconcentra. followed by oxyanions dependent on the major tated the result is that further evaporation will chemical controls in the pit lake. Where pit lake chemistry shows 2maca > m alkalinity . Rees 1998). water chemistry is generally Na-SO4-Cl. circumneutral and oxyanions are strongly tion can be predicted to occur over time. Eugster & Hardie differing evapoconcentration paths would be 1978. Generally. This is the probably the most important control Concentration processes for MPL water quality (Fig. In this scenario. Such conditions are observed in several natural lakes such as Pyramid Lake. On the adsorbed and thus are not accumulated in the basis of this approach. As binary salts are precipi. where istry in a closed basin (Eugster & Hardie 1978). As most pit lakes are circum. precipitation and Ca concentrations increase with evaporation. Where the pit lake initially has 2mCa < malkalinity (here alkalinity is the sum of mHco3 + mCo3) then with continued evaporation. The influence of neutral Ca–Na–SO 4 –Cl chemistry. Eary (1998) demonstrated that two (Garrels & Mackenzie 1967. 10. Where 2mCa < result in an increase in the relative dissolved m alkalinity then a Na–HCO 3 –CO 3 SO4 concentration of the ion or molecule present in type chemistry will result and pH can evolve to greater concentration. On the basis of most case studies. On the basis evapoconcentration on brines is well understood of this. In very evaporate until CaCO3 becomes saturated and acidic pit lakes the major control will relate to . pit lake chemistry influence on MPL. The next 'chemical divide'. Eary (1998) assessed the pit lake. where evaporation exceeds in Nevada. accumulate over time.

12 0.01 5.29 0.35 0.4 < 0.32 0.6 10.06 < 0.02 < 0.22 3.5 11 10.03 1.074 0. 0–10 ft) Parameter* pH 7. alkalinity.02 < 0.7 18.2 0.5 295 295 271 649 342 265 384 555 625 738 680 531 814 856 566 620 590 789 867 747 858 870 851 955 1717 Alk. TDS.96 3.5 1 0.3 7.8 21.7 0.02 6.5 0.053 0.4 Cu ( m g l ') <0.05 0.086 0.58 7.76 6.08 3.97 8.1 5.2 26.18 3.37 As ( m g l ') 0.49 0.02 0.2 4.02 1. nd.02 1.74 1.29 0.3 0.48 0.74 0.1 3.92 7 5.48 3.02 < 0. 1 5 0.5 0.05 0.52 0.02 0.09 0.7 2.53 0.7 7.02 < 0.7 1.044 0. 1 5 0.02 < 0.02 < 0.29 6.11 Ca ( m g l ') 52.06 7.04 nd Nd nd 3.01 0.7 7.53 9.69 0.67 3.36 8.02 < 0.021 0.32 0.019 0.02 < 0.06 0.82 7.23 7.12 0.07 0.86 7.5 < 0. . Hydrogeochemistry of Summer Camp Pit (depth.7 47 49.04 0.9 5. total dissolved solids.4 0.41 0.4 nd nd Nd nd Nd nd nd nd nd nd nd nd nd nd nd 99 nd Nd 164 360 190 180 nd nd nd Mn ( m g l ') nd nd nd Nd nd Nd nd nd nd nd nd nd nd nd nd nd 0.4 1.81 0.93 TDS ( m g l ') 265.27 1.2 7.35 5.06 0.8 13 13. not determined.56 0.22 < 0.71 0.2 7.03 0.3 1 0.2 1 1.1 2.Table 7. 1 3 0.2 7.023 0.02 < 0.02 < 0.6 14 17.05 nd nd nd * Alk.4 10 1..11 0.47 0.03 0.02 Zn (rngl ') <0.2 3. ( m g l ') 149 155 187 III 0 109 160 93 0 0 0 0 0 0 0 79 14 6 64 68 51 58 64 30 81 80 Sulphate 37 97 nd Nd nd Nd nd nd 196 209 588 560 317 572 522 386 360 488 400 450 430 490 430 430 510 1000 (mgl – ' ) Fe ( m g l –') nd nd nd Nd nd Nd nd nd 3.

so these in turn are related to the abun- concentration processes. ization. a permanent hypolim- cularly in the former. Geological controls Another factor influencing the formation and Steady-state cycling characteristics of secondary minerals is the depo- Within an established MPL. This steady state may are related to acid generation and acid neutral- be imbalanced by seasonal effects or by evapo. Schematic diagram showing implications of hydrous ferric oxide chemistry on solute transport at different pH (from Deng & Stumm 1994). but parti. lake and. and buffer- change greatly. The speciation and dance of primary sources of these processes. sit geology in which they form. As the major ation processes will reach a steady state or geochemical reactions in a 'sulphide rock' mine equilibrium at some stage. then ing minerals. HYDROGEOCHEMICAL DYNAMICS OF MINE PIT LAKES 175 Fig. produce highly saline. dominantly calcite. the main the effects of seasonal turnover will become less controlling salts will be iron sulphates such as as turnover effects do not extend as deep into the melanterite or gypsum. highly acidic metal-rich brines (Nordstrom & Alpers I999b). 11. distribution of elements within the pit lake may namely sulphides. In either case. if a lake deepens. Over the last . release and attenu. particularly pyrite. Over time. sulphuric acid equilibrium. consequently. In the latter. evapoconcentration can nion becomes established.

The majority of porphyry pit lake results only a few are exploited by open-pit mining and reflect low metal. even where sulphides are present in lin et al. Shevenell et al 1999. Another sulphide zone that exists within occur on a large scale and. principally predicted. 5). understanding potential impacts anticipated from Where carbonates form a large portion of the mining sulphide-bearing mineral deposits (Fick. been partially oxidized and so have associated phidation epithermal systems and VMS deposits with the sulphides a series of Cu-Fe sulphate that host tens of per cent sulphides. Plumlee 1994. for example the Summitville of environmental geological models to provide high sulphidation system in Colorado (Gray et al some form of initial prediction mechanism for 1994). Du Bray 1995. non to low-acid generating. as such. relatively high concentration pit lake. Where they form a significant part buffering agents result in the more acidic metal. the silicified sulphide-bearing zones that have Thus. Typically. of the exposed pit then an acid pit lake can rich pit lakes (Fig. Nevada that tend to show low base distinct geological characteristics of these metal concentrations in the ore and a high pro- deposits (Figs 4-6). and are usually devoid of calcite or other as jarosite. 12. it can be observed that the high sul. but thus the data are biased towards copper porphyry some zones within porphyrys can produce a high deposits. and the presence and deposits. such pyrite. considerably higher and metal concentration Bowell et al 2000). Carlin-type gold deposits of the western potential for acid generation and metal leaching United States and volcanogenic massive sulphide (Bowell et al 2000). Relationship between Ca and sulphate in mass-balanced mine pit lakes showing the majority have as an upper control saturation with respect to gypsum. BOWELL Fig. calcite. aspects of metal deposits that influence MPL Although little data are available for pit lakes hydrogeochemistry will be the speciation and developed for sedimentary base metal sulphide proportion of sulphides. J. circumneutral low metal chemistry. 1992. few years several studies have examined the use significant impact. can cause porphyry is the chalcocite blanket or zone of . Those above were all below pH > 2 and have an iron sulphate - sulphuric acid control. These deposits also tend to prevail. what there is reflects as would be abundance of buffering materials. lower. On the basis of the above portion of calcite in the alteration zones related discussions.176 R. these are from (VMS) deposits (Price et al 1995). ore. A good example are the disseminated Mineral deposits do show a variation in metal sediment-hosted micron gold deposits of the and anion hydrogeochemistry related to the Carlin belt. dominantly salts and other sources of secondary acidity. it can be observed that the important to carbonate host rocks (Hofstra et al 1995). pH is Plumlee & Logsdon 1999. Owing to the nature of mineral deposits.

Geochemical trends over time for Summer Camp Pit lake. 1990-1998. 13. HYDROGEOCHEMICAL DYNAMICS OF MINE PIT LAKES 177 Fig. .

178 R. 13. (b) (Continued). . BOWELL Fig. J.

Tempel et al 1999. with a small pool these zones are present then metal leaching can developing in the west of the pit. chalcocite and covellite.05 mg l–1) in the pit Few detailed examples of pit lake evolution have sump. cuprite. 1990–1998. This had occur at relatively high pH (pH > 4–5).7H2O). Zn (up to 0. in response to hydrogen ion sulphate. 14.7–8.009rngl –1 ). . the period from pre-mining through to (Table 7). with As below the primary Camp Pit. except for the pit lakes on the 146 to 187 mg 1–1 (Fig. For one of these alkalinity in the sump and water pH was acidic pits. Se to the configuration of chalcocite. Possibly this reflects the increased solubility of HFO and stability of Fe3+ migration of oxygenated groundwater along the in solution.082 mgl – 1 ).3 to 8 Getchell Mine. Adsorption of arsenate likely in equilibrium with groundwater. suggesting that either After pumping was terminated at the end of pyrite oxidation was not appreciable or that operations in December 1991 alkalinity was Fig. Slightly elevated metal/metalloid several studies from 1995 to 1999 (Barta et al levels in the pool were recorded (Table 7). 1998). December 1991 (Fig. Pool alkalinity in March 1991 ranged from been published. chiefly arsenian pyrite. Owing measurable levels of As (up to 0. The acidic pH was accompanied cessation of mining and formation of a MPL has initially by a low TDS. it does not (up to 0. due to precipitation of been constantly monitored from 1990 to the evaporative salts like gypsum. 1998).3). minor jarosite and relict pyrite (Lowell & During pit development in 1990 water quality Guilbert 1970. Natural at mildly acidic pH (4-6) limits As hydro- groundwater has an alkaline Ca–Mg–HCO 3 morphic dispersion but at lower pH (< 3) arsenate chemistry with variable amounts of arsenic and can be mobilized. 13b). this time pH of water around the SCP was circumneutral to alkaline (pH = 7. halotrichite (Fe2+Al2(SO4)4. Nevada that have been subject to (Fig. (Fe2+SO4. At and consequently produces alkaline drainage. dependent on contact with gypsum and activity stabilizing arsenic acid (H3AsO4) and the sulphides (Table 4). submitted). Water Case study of pit lake chemistry: Summer chemistry shows low salt load and trace element concentrations. The pH is circumneutral. with sumption was sufficient to buffer generated H+.1 ) and TDS (360 mg 1–1). From June 1991 a gradual increase mineralized shear zone in which water reacted occurred towards the end of operations in with primary sulphides. 14). low produce protons on oxidation but hydroxyl ions sulphate (110mg1 .20- (SCP) area was undisturbed and as such was most H2O) (Bowell et al. By 1997. Saturation Indices for major minerals in Summer Camp Pit lake. this typically comprises groundwater alkalinity and host rock acid con- of goethite. 13a) and pH from 7. Shevenell & Pasternak May 1991 acid generation had exceeded available 2000. Nevada drinking water standard (0. Parshley et al.22- Prior to mining in 1990 the Summer Camp Pit H2O) and copiapite (Fe2+Fe34 (SO4)6(OH)2. If remained fairly constant.l2mgl– 1 ). HYDROGEOCHEMICAL DYNAMICS OF MINE PIT LAKES 179 supergene enrichment. melanterite present (Bowell et al. Guilbert & Lowell 1974).

180 R. pH fell from 7-7. metalimnion and hypolimnion (below 20m). by May 1992 decreased to 3. 13b). BOWELL Fig. so remained in the This increase in acid generation correlates to range pH 3-4 (Sverdrup 1990. Fe and As.5 to 6. 15. rather. 16.67 (Fig. Kwong & flushing of secondary acid salts as the water Ferguson 1997). water pH was buffered from May 1992 by due to increased rates of acid generation. sulphate. In The zero alkalinity observed is most probably the absence of excess carbonate alkalinity. steadily consumed until April 1992. . hydroxides and silicates. Density (sigma excess) plot with time for the SCP. not due to cessation of buffering but. During this period available alkalinity was consumed as acid gene. Fig.89 in April 1992 and ration exceeded alkalinity recharge (Table 7). Chemistry-time plots for the SCP showing depth variations in the values of pH. J. when all level recovers in the pit. Depths are set at various intervals to record changes within the epilimnion (above 9m).

This is most probably due and rainwater surface run-off. Copper Flat Pit lake. 13a). . Most probably this is because hydroxide. Sternberg lode. pH increased to circumneutral levels to increased acid production and be utilized in (Fig. 13a). 'reactive' pit walls were exposed and these probably due to adsorption onto ferric hydroxide contributed metals. SCP hypolimnion sediment. sulphate and acidity to the and clay minerals. (a) Section showing euhedral to subhedral crystals of mackinawite (Fe. HYDROGEOCHEMICAL DYNAMICS OF MINE PIT LAKES 181 The slow accumulation of dissolved iron Water was abstracted from SCP for oper- compared to sulphate is most probably due ational use from May 1992 to September 1992. New Mexico.7H2O) or ion flooded open pits and the underground operations exchange of Zn with other divalent cations elsewhere on the Getchell property (Fig. during clay mineral-water interactions. sphalerite (ZnS) and orpiment (As2S3) in kaolinite-quartz matrix. Hillsboro County. Note sphalerite and orpiment are most probably detritus grains. being essentially a Na–Ca–SO 4 type with the Fig.Ni. to a greater portion of iron being reprecipitated The high rate of acid generation was maintained as secondary minerals such as hydrous ferric during this period. (c) Backscatter image of orpiment euhedral crystal showing negligible corrosion. SCP hypolimnion sediment. chalcanthite and gypsum on wall- rock. Photomicrographs of sulphides from pit sediments and sulphates from pit walls. even ary mineral leaching due to water level recovery at neutral pH.Co. Following the addition of alkaline water into the sonite dissolution might occur in response SCP.Zn)9S8. SCP hypolimnion sediment. Smith. whilst mackinawite is probably authigenic. (d) Halotrichite. (b) Backscatter image of mackinawite euhedral crystal showing negligible corrosion. The low pH to secondary processes such as dissolution environment was mitigated in late 1992 with the of soluble zinc minerals like smithsonite diversion of HCO3-bearing water from other (ZnCO3) or goslarite (ZnSO4. with the water chemistry carbonate. Arsenic and selenium during this pumping maintained the water level such that period do not show significant increases. 17. Total dissolved solids have remained the buffering process along with calcite relatively constant. Zinc levels show a steady pit lake through sulphide oxidation and second- increase from the end of operations.

As can be observed for much of the zone. At the SCP an important external Fig. Turnover events such as these could be as pararealgar. 14). . This aeration comes oxygen and sulphide oxidation model reflects the internal and external processes occurs in the basal sediments. Similar processes will occur in other but they do occur within the pit wall-rock and thus deep pit lakes. Co. A the mine pit lake. model was developed to assist understanding of the basal layer is briefly aerated resulting initially the development and geochemical-limnological in higher pH and lower metals. processes have been observed in deep-sea anoxic Chemical variations in the pit lake over the sediments (Morse 1994) and natural wetlands same period reflect its stratified nature. 18. As and metals in the basal Utilizing all this information.182 R. bedrock geology. pH. as well Table 4). 18). turnover (as shown for February 1998 for SCP. sulphate-reducing bacteria have plots for the pit lake set at various depths been identified (Gannon et al. the nature of the the hypolimnion. three all show authigenic morphologies (Fig. sphalerite and pyrite. for example the Sleeper Mine in a detritus origin cannot be ruled out. stratification of this pit lake is currently being Further variations in pit lake water quality prepared (Parshley et al. Although sediments of the SCP through sulphidization in not discussed here in detail. On turnover. 16). reflect changes in the chemistry of pumped mine Attenuation of metals occurs in the basal water rather than pit lake processes. as nostic leach scheme used in this study involved a a permanent hypolimnion forms in the deep modified Tessier selective extraction procedure (20m) pit lake. Within this (Fig. In the hypolimnion of Summer stratified pit lake can be observed in the density Camp pit lake. J. and include mackinawite ([Fe. longer-term multi. Zn]9S8) and greigite (Fe2+Fe32+S4). Similar Nevada where the pit lake is over 60 m deep. The diag- magnitude of this turnover would become less. the basal layer is considerably denser than diagnostic leach studies to contain high metal the other two layers but in winter this changes concentrations associated with the reduced as turnover occurs (Fig. Over time the fraction (Bowell & Parshley 1999). high sulphate. but with the environment within the SCP lake (Fig. The latter both seasonal and. Rees 1998). BOWELL main mineralogical control being gypsum more detailed study of the limnological-depth saturation (Fig. possibly. Table 4) shows lower TDS water than pre. submitted). analysis the phases in this fraction have been turnover lake (such as the data shown for June confirmed as authigenic sulphides not observed in 1998. sediments have been found through year. Ni. This leads to low that control the water quality observed within pH and high metals as stratification reforms. From mineralogical that it acts as a 'self-cleansing' event and a post. a conceptual layer for much of the year (Fig. 15). A crucial impact of turnover is (Bowell & Parshley 1999). 15). 1996). Conceptual model developed for the Summer Camp Pit (SCP) based on the data collected in the field and by laboratory analysis to explain the complex interlinked limnological and geochemical mechanisms that control pit lake water quality. 17) year events. with low (Kwong & Van Stempvoort 1994.

lake hydrodynamics. BIGHAM.J.J.K.N. Conference on Acid Rock Drainage.A. 1994. Vancouver.. Geochemical controls on pit lake water quality. 1997. S. A. (eds) Environ- mental Geochemistry of Sulphide Mine Waste future water quality in mine pit lakes involve Mineralogical Association of Canada.I. Balkema..W. J. A. 1996. D. Mine Environment Neutral nating between historical sources of solutes in the Drainage Program. Applied Geochemistry. J. D. methods for predicting the JAMBOR. Ely. DAVIS. 9. California. Geochemical standing of the environmental assessment of pit and Limnological Assessment of Summer Camp Pit. Mineralogy of ochre deposits. O. J. pit wall run-off.B. 550. In: ALPERS. review and BOWELL. D.M. Seasonal variations of Zn/Cu ratios in Acid ALPERS. Nevada. C. & BYRNS.M. Prague. Fillipek... R. MANSA- support particularly J. improved the pits.V. R. & BLOWES. precipitation. Minerals.. L.M... Geochemistry of Sulphide Oxidation. 7. 1996. (eds) expressed within this paper are those of the author Geology and Ore Deposits 2000: The Great Basin and not SRK or its clients. Controls on ochre chemistry and This paper has made extensive use of field notes. Ottawa.A.J. PRICE. Pit lake water quality in lakes: observations and modeling using CE-QUAL. Limnologic conditions in three existing Nevada pit DAVIS. R.N.B. this practice has now ceased. NORDSTROM. R. In: IMWA Symposium Proceedings. J. BOWELL. LYONS. The views and opinions E. American DAVIS. Ottawa. 4. P. J.M. REES. 345–364. Secondary iron-sulphate ALPERS. Journal of IMWA.. B. Ayora in generation: geologic controls and baseline assess- reviewing this manuscript. Applied Geochemistry. A. 1997. In: Proceedings of Living with a Closed Basin. Refinement of such models and general under. R.A. evapor. American tal Geochemistry of Sulphide Oxidation. Rees. Reno. which resulted in a much ment. (eds) Environmen. R. A. J. D. C. J.C. (eds) Environmen- Mine Water from Iron Mountain. FUGE. Series. Discrimi- 31 May-6 June.M. May 2000. 1994. Reno/Sparks. D. 699–713. to reflect the major chemical and physical BIRD. VUORINEN. Geologic W. 96. 1998. J.L. In: CLUER.A. 98-102. Sadler. A review of the major processes influencing the BHATTI. Reno. London. R. GINGRICH. & MARTIN. BOWELL. & ASHENBERG. controls on pit lake water chemistry: Implications Williams. Most of these BOWELL... Connelly. KEMPTON. D. An processes operating within a MPL. M. T. and geochemical Waste '94.W. 1999. and I acknowledge the debt owed numerous September 1996. Johannes- manuscript through review and graciously approved burg. Nevada. A. hydroxides and oxides. P. water into the pit. & PARSHLEY. unpublished data and ideas held by SRK Con.H. J. of mineral deposit. NEWCOMB. Parshley. 1989. In: Tailings . C. particularly BOWELL. In: 1994. • limnology of the pit lake. Harding. In: Proceedings of the Fourth International genic trends. K. HARDYMAN. & SADLER. STRUHSACKER.A. BARTA. J. C.V. M. precipitation in coal and metal mine drainage.. W. Geo- logical Society of Nevada. S. Nevada. chemistry. D. W. & PARSHLEY. sulting. USA. Dey and L. & BLOWES.J. In: tal Geochemistry of Sulphide Oxidation. Getchell Gold Corporation. Canada. feldspars associated with the microbiological The major controls on MPL can be related to: oxidation of pyrrhotite and pyrite.L. Washington DC. In: Proceedings of Tailings and Mine ation. In: At the present time. Butte. W. lakes will be improved through characterizing Nevada.J. American ALPERS. Finally. minerals as sources of sulphate and acidity. & MORRIS.H.. The aqueous Chemical Society.M. the conditions in existing MPL. BIGHAM.G.W. 31–40. 90–107. V. & MASANARES. colleagues past and present for discussion. & TUOVINEN.J. & BLOWES. RSA. References CRAVOTTA.F. III 1994. Elko. BOWELL. Alteration of mica and hydrogeochemistry of MPLs has been presented. CONNELLY. R. 279-286. R. (eds) Environmental • geochemical processes within a pit lake. 291-323. Hockley.K. 2000. NARES.J. 799–823. 23-36. Montana. J.. & EARY. HYDROGEOCHEMICAL DYNAMICS OF MINE PIT LAKES 183 control is the discharge of underground mine BARTA. I acknowledge the Geochemical prediction of metal leaching and acid time and effort taken by L. Arsenic sorption by Iron oxy- methods are still in the process of development. Chemical Society Symposium Series. T. 550. 375-386. In: reports. Day. R. American • geology of the exposed wall-rocks and type Chemical Society. J. & PARSHLEY. SRK Report to Getchell Gold Corporation. & THOMPSON. 1998. Minerals Engineering. American Chemical Society. G. improved final version.. Robinson Mining District. Metals and the Environment II. C. 324-345. wall-rock applied to modeling post-mining pit water geo- leaching.. C. equilibrium and speciation. Nevada Water Association. B.M. EC. Washington DC.E. ATKINS.K. J. C.W. I acknowledge the for the assessment of water quality in Inactive open support of clients who provoked thought. Rotterdam. & MILLER. October Conclusions 1997. D. 1994. 550. and Beyond: Geological Society of Nevada Sym- posium Proceedings.. publication of the data within this paper. the western United States: An analysis of chemo- W2. 103–131. 1994. These major assessment of hydrogeochemical computer codes processes include groundwater inflow. Washington DC. Nevada Water Association Annual Meeting. hypothetical mechanistic and empirical models Canada. & BLOWES. D. Chemical Society Symposium Series. geochemistry of the Berkeley Pit. American Chemical Society Symposium. Lovgren and C.

1997.J. Oxidation. LEVENTHAL.. 1994. 1970. 550. Rotterdam.F.E. American Chemical Society Sym- American Chemical Society.. West Shasta mining district. Sediment-hosted Au deposits. The Geochemistry of Natural minerals.. J.). YT. 393-403. 1996.S. Mine Environment Neutral Drainage Program. vertical mineralization zoning in porphyry ore DC 67.J.H. Mineralogi- and sediments of West Squaw. & LOWELL. Geochemical classification of ceedings.M. 235-243.B. Vancou- 21. 388-396. B. LYONS. G. Acid production from sulphide minerals centration on water quality in mine pit lakes. J. In: Fourth Inter- FLICKLIN.S.J. 184-192.M. (eds) Environmental Geochemistry of Sulphide American Chemical Society Symposium Series. (ed. 89. R. L. & VAN STEMPVOORT. for point of zero charge for some common DREVER. S. Environmental Geology of pyrite in low temperature acidic solutions: Rate the Summitville. 1974.E. American Chemical Society. using hydrogen peroxide weathering. McAooo. 1995. John Wiley. & INSKEEP. N. Canada.184 R. Reactivity of aquatic and de watering. Environmental Science & Tech.T. & BLOWES. Aqueous Environmental Geo- TRO. BOWELL and Mine Waste '98.L. PEARCE. . & MACKENZIE. Geology. Geochemical and equilibrium trends in KARAMENKO. Montana. Utah. cal changes during NP determinations and their nology.I.) Equilibrium Concepts in Ottawa. A. Wasington DC. deposits. International Geo- EUGSTER.) Preliminary Compilation of Models.B.D. D. cycling of iron in natural waters. 435-447. posium. D. 1994. Golconda.D. (ed.H. 223–236. DOLLHOPF. Geochemistry.T.A. 1988.W. 1988. M.H. ALPERS. G. Saline lakes. J. Surface United states Geological survey. W. W. MOORE. Y. A. M. 223-242. PLUMLEE. R. N. 1996. H. 30. FUGE. W. Canada 1729–1744. the chemical composition of some springs and Mine Environment Neutral Drainage Program. C. R.. & STUMM. H..M. KRANZ. D. New Jersey.. 64.W. A.. 1509-1520.A.. In: logical Reviews.. A. Natural Water Systems. Washington. posium Series.E. & WANG.W. Englewood Cliffs. Chemistry.. US Geological Survey Open File Report.. & HARDIE.I. Oxidation of ATKINSON. California. 99–109. L. J. Prentice-Hall. 2nd ed. J. Mine Environment Neutral mine drainage and natural drainage in mineralized Drainage Program. Prentice-Hall.. 1997. A. 67.D. D.K.M. SMITH. & GUILBERT. F. G. K. Preliminary Compilation HERAN. Acid mine drainage in Wales and KWONG. & MEIKLE. Applied Hydrogeology. & Conference. GRAY.BLISS. COOLBAUGH. GANNON. Advances in Chemistry LOWELL.R. 65. Geochimica et Cosmo- 2006-2014. Calculated values Geochemistry. HOFSTRA. E.R. mental Geochemistry of Sulphide Oxidation.S. A. 2000. Physics. 9. & WORTHINGTON. 261–274. L. Applied HIEMSTRA & VAN RIEMSDUK 1996. laws and surface textures. & BARNES. 382-292. ver. (eds) Environ..D. Models. Mine lake geochemical prediction 381-384.F.W.M. L. 1992. NORDSTROM. New Investigation of the Sulphate Reduction Potential Jersey. 271-281. Complexation Modeling — Hydrous Ferric Oxide. Economic Geology. implications.M. 1997. 1978. New Jersey. D. Vancouver. V O L B E R D I N G . Report to Getchell LAWRENCE. PEARCE. & FICKLIN. & FERGUSON. LANGMUIR. R.A.B. GUILBERT. Determination of Gold Corporation.Probablis- FETTER. DZOMBAK. JENNINGS. Washington DC. Ottawa.W. J. Proceedings ICARD '97. EARY.. neutralization in the prediction of acid rock GARRELS. (e. 237-293. (ed. vol 550. L. 478-549.&MALEY. 1997. Vancouver. Interaction of acid mine draiange with waters KWONG. LERMAN. 118–123A. 1998. In: Proceedings ICARD '97. Applied Geochemistry. J. C. Attenuation of acid rock drainage in a natural In: ALPERS.P. 451–464. 11.. T. W. wetland system. D. W. lakes. Descriptive Geoenvironmental Mineral Deposit 95-831. K. Ottawa. & DAVIS. 179. Billings. 1999.A. in the Summer Camp Pit Lake. Wasington DC. Economic Geological.L. & BLOWES.N. Mine pit lake. 3rd edn. W. Englewood Cliffs. J.. 1994. 1986. US Geological Survey Open File Report. Field chemistry. J. Lakes. KEMPTON. F. 1967.J. 1994.M. LOCKE. P. Colorado.. D.. Lateral and Series. W. A simple 631-638.P. of Descriptive Geoenvironmental Mineral Deposit Du BRAY. HAVIS. Origin of drainage.H.963–988.. for the SPJV MacDonald Project. & national Conference on Acid Rock Drainage Pro- MCHUGH. D. Biogenic factors in the mine pit lakes. Journal of Colloid and Interface Science. J.Y 1969. Balkema. Environ- Institute of Mining Bulletin. & MOREL. GOULD. American Chemical Society. Waters. S. SCHAFER. 488-508. Candian standing the water quality of pit lakes. NICHOLSON. WIELINGA. & MCKIBBEN. A. J. 1997. A. 95-831. In: Proceedings of the Seventh Water-Rock KIRK. New York..A. chimica Acta. W. Y. In: Billings Sym- FILLIPEK. New York. Englewood Cliffs.G. Rotterdam. American Chemical Society. areas.implications for redox Waste '97. EARY. Under- zoning patterns in porphyry ore deposits.H. Springer. 1995. 14. 373-408. 1994.L. BENNETT..J. BC. Wasington DC. tic prediction of water quality in the Twin Creeks Prentice-Hall. DJ. J. GRIMES. influence of ochre precipitation on water chemistry. mental Science & Technology. & PERKINS. Balkema. W. Variations in MILLER. & Du BRAY. Rotterdam. S. R. F.A. Applied Journal of Geochemical Exploration. J. 23-36.. YT. 15.. In: Proceedings of Tailings and Mine iron (III) oxyhydroxides .A. 1990.S. PLUMBLEE.d). 50. model for the management of mine pit-lake filling DENG.D.A. Predicting the effects of evapocon. Balkema. Canada. 1996. POLICAS. Nevada. E. L.P. W.M. formation of sedimentary ores.

BERG. PARFAIT. 289-297.G. 227-264.. 4. YOUNGER. Lund Uni- 33455-33462. 51. J. K. & ALPERS. PLUMLEE. Negative Colloidal Surface. Lund. of. D.. 1983.M. & HENRY.. & EDMUNDS. & M. Chemical Society.B. pH. R. Rotterdam. Chemical properties of variable TEMPEL. A. can Chemical Society Symposium Series. CONNERS.R. 669-687. D. K. D. Oxide Zone Geochemistry.G. The interaction of anions Economic Geology. HYDROGEOCHEMICAL DYNAMICS OF MINE PIT LAKES 185 MORSE. The Kinetics of Base Cation National Academy of Sciences USA. D.. Processes con. Processes Operating in the Whitworth No.G. E. Geochem. RJ. Mineralogical Society of Nevada Bureau of Mines and Geology Open File America.. trolling acid attenuation in an unsaturated zone of a SCOTT.G. 1987. 96. Report. Australia.. 7. Chemistry of the Solid-Water environmental restoration at the Iron Mountain Interface. PL.O.M. California. Cosmochimica Acta. Chemical. New York. H. Sussex. limnic eruptions? In: Tailings and Mine Waste '97. J.12. DYSINGER. Vol 95–4. & L.D. 167–194. L. SHEVENELL. E. (eds). Stratiform Ore Deposits. W. 1990. L.N. weathering. waters flowing from abandoned coal workings in FIELDS. C. . Amsterdam MORRIS. DRIESNER. & BLOWES.D. Australia. 15. DOYLE.K. RUNNELLS... M.D. chemistry of Mineral Deposits. 1994.I. 1995. Balkema. M.H. minerals. In: trochemistry. L. Balkema. John Wiley. W. D. The kinetics of calcium carbonate and Abandoned Mines in Nevada. Regolith mineral deposits. 1987. Special Publication. 13. Elsevier.K. MITMAN.J. F. W. J. Water Quality at Inactive ing Geology...W.L. Gossans. 11. Release of toxic metals via oxidation REES. 1987. PRICE. 13. WOLF. 1997.L.. SHEVENELL. 529-542. mental Geochemistry of Mineral Deposits. 869–871. 475-492. 133-160. 1980.J. Exploration Geochemistry in Tropical Terrains. Rotterdam. L. HENRY.H. LOGSDON.D.H.. S101–S113. 1995. 1994. CHATHAM. J. 15. & pit mines in Nevada. efflorescent mineralogy and consequences for STUMM. W.. D. C. J. Post-mining pit lakes: Predicting pentlandite and violarite. New York. MILLS. J. R. Chemical Geology. NORDSTROM. D.. & chemical stability of the nickel–iron sulphides YOUNG.J. R..A. NORDSTROM. A.L. versus observed filling rates of the historical NICKEL.F. 1–14. J.N. In: THENG. D.l ALPERS. 1992. LECHLER. STRUHSACKER. METESH. THORNBER. New York). B. & YOUNG. J. In: Getchell Pit Lakes... & FILLIPEK. Longevity of the Treatment of authigenic pyrite in resuspended sediments.. W. (eds) Environ. Nevada.A. 2000.E. 1982. Nevada. May 2000. 2000: The Great Basin and Beyond: Geological NORDSTROM. 1997.. THORNBER. New Zealand Soil Science predicting arsenic concentrations in a mine pit lake.G. C.S. and weak acids with the hydrous goethite surface. S. Tailings and Mine Waste '96. DESILETS. (eds) Handbook of Stratabound and J. TWIDWELL. Modeled A. 1975. PLUMLEE. Geochemical modeling approach to Variable Charge. Queensland. & County Durham. L. 6A.S. Applied Geochemistry. (eds) The Environ.B. 6A. W.. M. G. J. & PASTERNAK. Environmental geology models of Handbook of Exploration Geochemistry.A. PRICE. 1561-1571.. W. R. K.. DIEBOLD. 469-478. nickel sulphide deposits at Kambalda.W. A. & ALPERS. NORDSTROM.. 101–117. P. & charge soils. 1998. THORNBER.. DUAIME. and classification Triassic sandstone aquifer UK. 1985. R. RJ. SHEVENELL. Butte.K. In: BUTT. T. 1995. 1978. Are pit lakes susceptible to ologist.M. In: REEDER. P.V. H. 65-99.) Cooperative Project Amongst State Agencies. 1999a.. physical and disslved oxygen and by ferric iron. Tailings and Mine Waste '97. & DANIELS.M. A.G. 261-390. RIGBY. Reno/ and the consequent formation of secondary Sparks. 5-6. R. C. Constructed Wetland for Amelioration of Acid mental Geochemistry of Sulphide Oxidation. Reviews in SIGG. Mine Water & The Environment.R. 543-548. 1990. A. Mineralogical and electro- PlLLARD. Wasington DC 550. 28. Aqueous pyrite oxidation by H.L. M. Ellis in Economic Geology. Geochemistry and limnology of Summer sulphides.R. W. Applied Geochemistry.. CLUER. Applied Geochemistry.H. K.. American ROBINS. SHEVENELL. Amsterdam Vol. versity Press. Soil Reno. LOGSDON. C. Montana. 2nd edn. Ameri.. Balkema.R.U. L. gossan-derived members of the alunite-jarosite bonate minerals.K. American Miner- MURRAY. Geochimica et biological interaction at the Berkeley Pit. 178–187. Exploration and Mining Geology.. 253–267.K. LECHLER. 2.E. MORSE. G. I: a chemical model based on massive Camp Pit. (Wiley-Interface. Aqueous pyrite oxidation Society of Nevada Symposium Proceedings.A. in the absence of car. (eds) The Environmental Geo. Release due to Chemical Weathering.. Horwood.) Soils With PRICE. (eds) Geology and Ore Deposits Vol. ZEIGERS. 14. 1983.K. C. In: PLUMLEE.M. C.. C. G. Moss. J. PA.C. G.A.A.J. 1999.L. Reviews in Mineralogy. M. Solid solution in. Chemical processes during Rotterdam. 16. & STUMM. HUANG. Society.. Quarterley Journal of Engineer- LOMBARDO. (ed. Nevada. PARSHLEY. family. BOWELL. SEG Newsletter. L. Controls on pit lake water quality at sixteen open istry of acid mine waters.. Superfund site.W. C. Science Society of America. Geological Society of Nevada. Elsevier. (ed. HARDYMAN. Report of a dissolution and precipitation. DURBIN.S. 1992. Reviews WILLIAMS. Supergene alteration of Submitted. T. P. 1999. Proceedings of the SVERDRUP.573–-583.G. Western PETERS. K.N.K. Mine Drainage MSc thesis University of Leeds. A. 1992. Hydrogeochemistry of mine- CHRISTENSEN. I999b. Journal of Applied Elec- lake chemistry and assessing ecological risks. MOSES. HERMAN. In: Acid Sulphate Weathering. J..

.

198. a flow model permitted a detailed description of the behaviour of the column at a daily time- scale. and extending 45 km down. which. silicate minerals dissolved. sludge is expected. Moreover.es. 08034 Barcelona. even for gravels or low reactive sand. impeding the dissolution of silicate minerals and precipitating amorphous Fe(OH)3 in the place of jarosite.S.csic. 08028 Barcelona. C/Jordi Girona 1-3.carrera@upc. 1). with minor amounts of et al. The sandy and clayey columns behaved very differently. In the clayey column. As a et al 2001). According to the predictions the dissolution of jarosite was very important to maintaining the pH at a value of approximately 2. The model also permitted rate laws proposed in the literature for pyrite oxidation to be discussed.es) Abstract: Within the framework of a study on the impact of a mine-tailing spill at Aznalcollar. we investigated the oxidation of pyrite and other sulphides by means of two column experiments and reactive transport modelling. The sulphides stream from the impoundment (Lopez-Pamo were mainly pyrite.csic. were explained by their coprecipitation in the jarosite. & ROBINS. N. Modelling the oxidation of sulphides in an unsaturated soil MAARTEN W. CARLOS AYORA2 & JESUS CARRERA1 Department of Geotechnical Engineering and Geosciences. 187 Geological Society. with a d50 value ranging sludge in the tailing dam were retained in the between 4. C/ Lluis Sole i Sabaris s/n. in its turn. sulphides with some silicates. S W Spain. galena and chalcopyrite. From: YOUNGER. The columns were located outdoors for 15 months and leached 10 times. uppermost layer of the soil.L. We found that the oxidation of pyrite by Fe3+ was not faster than by O2.5 and 12 um. The columns were filled with pyritic sludge mixed up with a sandy and a clayey soil. the tailing dam of the following the clean-up work.saaltink@upc. 0305-8719/02/$15.00 © The Geological Society of London 2002. The sludge was composed of to 1 m in thickness.05 m are predominent. Technical University of Catalonia (UPC). the model was used to predict the behaviour of other soil types and other sludge contents. respectively. however. As a result. it is possible that some sludge acid drainage and metal pollution of the river (from 0. (eds) 2002. SAALTINK1. approximately sphalerite. Prior to simulating reactive transport. This parameter controlled the amount of O2 that could diffuse into the soil. Building D-2. Spain (e-mail: maarten. Spain (e-mail: cdomenec@ija. P. thick in the Agrio Valley and the upper part of banks and adjacent farmlands were covered with the Guadiamar Valley. Finally. Special Publications. approximately 2600 ha of river. The images Aznalcollar mine (SW Spain) collapsed and the confirmed the presence of sludge in the soil valleys of the Agrio and Guadiamar rivers were with an irregular distribution (Anton-Pacheco flooded with 4 X 106 m3 of sulphide sludge. On 25 April 1998. where gravels and sands a blanket of sulphide mud ranging from 0. jesus. Given that the upper- channelled lower course of the Guadiamar River. The oxidation will result in Consequently. cayora@ija.es. CSIC. 187-204. In the sandy column. The most important parameter extracted was the hydraulic saturation. The high concentration of Zn in the leachates was consistent with the concentrations predicted from sphalerite oxidation. The grain 2X 106 m3 of acidic water coexisting with the size was very fine. most part of the soil remains unsaturated for Most of the sludge was removed using heavy much of the time. oxidation of the remaining machinery in the 4 months after the accident. contrary reports in the literature. .1 to 5wt%) remained mixed with the and ground water. 1999) (Fig. London. The low As and Pb concentrations. the dissolution of dolomite kept the pH high.es) 2 Institute of Earth Sciences Jaume Almera. pH dropped due to the oxidation of pyrite. Mine Water Hydrogeology and Geochemistry. Sludge deposition was particularly consequence. providing Na and/or K that precipitate together with Fe and SO4 as jarosite. affected the rate of pyrite oxidation. CRISTINA DOMENECH2. A number of The objective of the present work is to predict airborne multispectral scanner flights were the evolution of the pore water chemistry made to map the sludge remaining in the soil following the weathering of the sulphides.

following the method des- is able to model reactive transport in a saturated cribed in Querol et al. oxygen diffusion through the 1976). A realistic estimate of Fe3+ concentration would be also of interest. Mayer et al. A number of reactive preserved. the soil samples were In order to predict the evolution of the water in leached with distilled water. 2000) developed the code MIN3P. In the case of soils. is not assumed. 0. (\994a. we predict the behav- In previous papers. Among them. 2). which was applied to and formed a layer of about 0. Wunderly et al. Mineral-water reactions are described by means of kinetics laws and no local equilibrium is assumed. the kinetics of sulphide iour of the pore water chemistry for different dissolution and the behaviour of each contami. SAALTINK ET AL. using Cu radiation and graphite monochromator. the diffusion of oxygen is the main parameter governing the dissolution of sulphides and Column experiment release of metals. which is commonly assumed in conventional geochemical calculations. may not be true in nature (Nordstrom & Alpers 1999) and hence. were filled minerals. Another two columns with soil and coupled the code PYROX.2 m of mixed soil saturated aquifers. wm. It was observed that affected by the Aznalcollar accident. which coprecipitate some of the with two mixtures of soil (90wt%) and sludge contaminants. A realistic modelling of silicates is required because they may buffer the pH of the system and supply the cations required for secondary phases to precipitate. The mixture was the pores of the soil. scenarios of sludge-soil ratios and mineral nant present in the sludge has been investigated associations. heterogeneity on the water quality (Gerke et al. fraction was estimated from the net intensity of . (10 wt%). Location of the Aznalcollar mine and of the First. oxygen diffusion and pyrite oxidation in an A layer of quartz sand was deposited at the unsaturated zone. Walter et al. More recently. which fluorhydric acids. the mass minerals and precipitation of secondary phases. dissolution of soil found in Table 1. however. our model includes the The determination of the clay minerals required basic processes required to describe the reactions additional treatments of the sample with glycerol taking place in an unsaturated zone: multiphase and with a temperature up to 550°C (Thorez flow. as Fe3+ along with O2 can oxidize pyrite as well as precipitate in secondary minerals. b). such as those expected in the soil (Domenech et al 200la. These codes Alastuey et al. The chemical composition of the sludge used 1998) and to model the drainage of a mine in the columns was that of sample PM1 of impoundment (Bain et al. whereas the variation in pH is responsible for the precipitation of secondary Two columns. affected by the spill.085m in diameter. The mineralogy or an unsaturated zone without assuming local was determined by X-ray diffraction (XRD). (1996) and sludge. This coupled bottom of each column in order to facilitate the code was applied to study the effects of spatial drainage (Fig. The two soil samples were regarded as transport modelling studies have been applied to being similar to the upper layer of the soils pyrite oxidation and acid mine discharges. This is especially important for silicates. sulphide oxidation. (1999). 2000). to MINTRAN. W. b) described the heavy machinery destroyed the soil structure the code MINTRAN. heat transport. which described the without sludge were also prepared as a blank. (1996). which are reputed to be in disequili- brium with surface water. is soil pores. Similarly to MIN3P. of the samples in a mixture of nitric and (1999. Prior to mixing. we use the model to simulate and discuss the results of the experiments. 1. The chemical composition assumed local equilibrium for all the chemical of the soil was obtained from the total digestion reactions. The mass fraction of each mineral. performed under constrained conditions. a reactive transport model made mechanically so that no structure was has been constructed. Then. Fig. Equilibrium between aqueous redox species such as Fe2+ and Fe3+. Once the main processes are modelled.188 M. equilibrium between minerals and water. we briefly describe the experiments area affected by spill of sludge (rectangle). During the cleaning works.

5 Chlorite 5 7 35 Hornblende 9 Gypsum 6 1 Dolomite 10 Pyrite 73 Sphalerite 1. chalcopyrite. K. The concentrations of As. the lack of O2 will inhibit the oxidation element i and vm. Mg sulphates and hydroxides precipitated during Table 1.6 . Pb. gypsum. as a result. 1999). The concentration of Na. Mm is the molecular mass of gas and less water. Al. during wet the mineral m. The most according to: important process that brings the O2 from the atmosphere to the soil is the diffusion of oxygen in the gas phase. The leachates were recovered. galena. sphalerite. with the exception of one particularly intensive rain shower. the oxidation of sulphides and. and were leached 10 times with 100– 250cm3 of Millipore MQ water.5 K-Feldspar 22 Illite 5 7 35 1. Mi is the atomic mass of the weather. filtered and analysed. OXIDATION OF SULPHIDES IN UNSATURATED SOIL 189 and K were assumed to be entirely contained in the pyrite. Ca. Fe2+. the more important peaks of the X-ray diffracto- grams (reference intensity method) and mixtures Modelling of pure minerals for calibration. The results of the analyses of Fig. Rainfall was generally not intensive enough to produce a leachate. 2. the leaching solutions are given in Tables 2 and 3. Zn. Fe + and total Fe concentrations were determined by colorime- try (To et al 1999). chlorite and illite. In hydroxides. of pyrite (Nordstrom & Alpers 1999). Fe and Zn in solution was measured by means of ICP-AES (inductively coupled plasma-atomic emission spectrometry).4 Galena 0. the precipitation of sulphates and hydroxides will occur preferentially where Xi is the concentration of the element i in during dry weather when the soil contains more the analysis (ppm). Ca. On the other hand. This is equi- valent to 15-45 mm of rain. Mass fraction (w m ) of each mineral in the sludge and in the soil used in the column experiments Sludge Sandy soil Clayey soil Quartz sand wm X 100 wm X 100 cum X 100 wm X 100 Quartz 8 31 30 98 Albite 23 10 0. Therefore. In the case of the sludge. while rainwater will dissolve the the calculation. respectively. all the Fe.i is the stoichiometric of sulphides and the precipitation of sulphates and coefficient of the element i in the mineral m. Sketch of the columns used in the experiments.8 Chalcopyrite 0. the mass fraction was calculated from the O2 is the ultimate driving force for the oxidation chemical composition (Alastuey et al. Cu. The columns were located outdoors on the roof of the laboratory in Barcelona for 15 months. Cu and Pb were determined by ICP-MS (inductively coupled plasma-mass spectrometry). Mg.

6 6.4 0. . Chemical composition of the leachates of the column of sandy soil and sludge.1 2. — .05 <0.0 2. 605 630 660 .Table 2.4 3.3 3. .8 5. 5 567 2 153 1 909 312 723 Z n ( m g l ') 1 358 427 285 361 913 376 714 225 491 135 105 Cu (mgl ') 0. 200 200 200 Vcollected (CHI3)* 52 29 63 65 93 102 122 116 122 134 126 72 pH 7.4 4.1 0.05 5 17 17 As(. 679 . see equation (3).1 Eh (mV) .004 1 1 4 42 35 108 34 82 25 22 Pb(ugl–') 0.05 <0. .05 0.3 1.6 K(mgr –') 7 11 10 10 8 6 <5 <10 .05 7 792 83 <0.0 2. Sample 9 is taken after an intensive rain-shower.ugl ') 5 9 46 26 29 126 2677 36521 19599 68678 14 120 19494 Vis a volume. . - Ca(mgl ') 46 752 494 470 436 530 377 632 290 520 216 426 Mg(mgl –') 6 255 247 134 148 271 194 797 283 640 175 160 Na(mgr –') 25 16 11 7 7 12 1.4 4. .5 <10 <10 <10 0.1 3 3 19 379 517 1581 578 1593 424 361 sor(mg1 ') 27 3519 3 101 2312 2486 6753 7260 28365 10757 30972 8262 12669 Fetotal (mgl – ') 0.05 <0.1 4 0. <10 <10 <10 Al(mgl–') – 0. .8 <0. .8 8 227 914 5891 2 153 7 369 1 714 3088 Fe 2 + (mgl ') .4 2. .5 2. the others after leaching events 0 1 2 3 5 6 7 8 9 10 11 12 Date 2/03/99 2/03/99 15/03/99 22/03/99 6/04/99 26/04/99 21/06/99 10/09/99 15/09/99 9/12/99 7/03/00 5/06/00 t (days) 0 1 14 21 36 56 112 193 198 283 372 462 Vadded (CHI3)* 100 200 200 200 200 200 250 200 . .

5 7.1 <0.1 0.0 7.laoie 3.1 <0. 67 43 98 37 24 1 . .8 6. me omenafter teaching\ing events 0 1 2 3 5 6 1 8 9 10 11 12 Date 2/03/99 2/03/99 15/03/99 22/03/99 6/04/99 26/04/99 21/06/99 10/09/99 15/09/99 9/12/99 7/03/00 5/06/00 t (days) 0 1 14 21 36 56 112 193 198 283 372 462 Kdded (cm3)* 100 250 200 200 300 350 400 300 .6 <0. . .1 <0.9 7. 624 . see equation (3).7 A s ( u g l – ') 7 20 16 28 25 20 21 28 59 14 10 12 * Vis volume.3 7.1 0.8 Kdngl-1) 15 14 8 10 14 16 13 27 17 <10 7 3 HCO3– (mgl – 1 ) . - Ca(mg1 – 1 ) 36 660 530 610 660 630 650 890 630 600 610 120 Mg(mg1 – l ) 5 95 130 160 200 210 190 300 230 210 230 41 Na(mgl – 1 ) 22 17 9 10 16 22 22 35 20 16 16 2.1 <0. Chemicai! composinoion of the teacnatesofthecolumn oiumnofclayeyey soil and ssludge.5 0.6 6. 645 626 629 .9 <0.1 <0. . 49 . . Samplee 9 is taken asner an imensi ve rain-snower er.8 6.0 8. 85 SO24– (mg1–1 24 2460 1 740 2 100 2280 2250 2 100 3300 2430 2220 2340 450 Al(ug1– 1 ) 74 140 <30 <30 130 300 520 67 1 100 380 <30 <30 Fe ( u g 1 1 ) 130 37 57 31 28 140 100 61 4200 420 12 26 Z n ( u g l – ') 330 310 800 370 1 100 900 55 100 580 800 320 21 cu(ug1 -11 ) 23 30 20 15 35 50 34 46 87 60 71 7 Pb(ug1– ) 0.8 Eh (mV) . 300 350 350 Vcollected (Cm3 )* 48 29 91 21 71 96 79 75 Ill 114 100 90 pH 7.1 1.8 8.1 7.6 7. .

It calculates pressure of gas and liquid. the volume of water per volume and convective heat transfer between the of pores) and capillary pressure (i. by Carrera et al. to describe the complete process expressing the flow of liquid. More information about the CODE- • flow of gas due to gas pressure gradient BRIGHT code can be obtained in Olivella et al. hydraulic saturation and temperature from soil properties (porosity. Transient flow and heat transport model CODE-BRIGHT models flow of water and gas and transport of heat. heat exchange with and A are those that better fit the experimental values the atmosphere. As no then used to establish the conditions for the data on this parameter were available. Retention curves relate hydraulic cal processes (solar radiation and advective saturation (i. hydraulic satur. 3. dis. CODE-BRIGHT (Olivella et al. Another input parameter for CODE- ation and temperature). pressure of soil surface and the atmosphere). gas phase. etc. The fine of the transient flow and heat model was to study grains of the sludge (silt) make the saturation the evolution of the flow and heat parameters of curve of the sandy soil close to that of the clayey the column experiments (flow. retention curve. W.) are simulated as boundary (dots. Saturation curves for the mixture of clay and perature. and a code for reactive transport. oxidation-reduction. and relative and intrinsic permeability). which consisted of Therefore. e. The best fits were for flow and heat.e. This also includes meteorologi. SAALTINK ET AL the dry season and will transport the solutes to conditions. We adopted the same approach used the run-off and groundwater. This also in. 1996). The procedure was a simplification of the real process calculated evolution of the hydraulic saturation and it was imposed by the present state of was compared with the experimental moisture development of the codes. The objective P0 and A are retention curve parameters. the pressure of gas and liquid. . gas minus pressure of liquid) for a given • transport of solutes by advection. 1998). 3). The values of P0 evaporation. quarsitic sand used in the modelling. (Dimos 1991) and fitted to the Van Genuchten • transport of heat both in the liquid and in the model (Van Genuchten 1978) (see equation 2. the mixture of sand and sludge and the velocity). (1996). the evolution of the columns with in the gas phase (vapour). rain. This tentative values from literature were used. RET- RASO (Saaltink et al. respectively. • transport of water in the liquid phase and In this study. humidity. method. soil. squares and rhombs) in equation (2) (see text). All equations are solved simultaneously • flow of liquid due to a liquid pressure by applying the Newton-Raphson iterative gradient. and state flow and reactive transport. (probably less important). dissol- ution-precipitation and surface sorption. sand and sludge and clay and sludge during a cludes meteorological processes such as period of outdoor exposure was modelled. For all these cal. Two types of models were developed: (1) transient flow and heat transport without solute where S is the hydraulic saturation. different steady-state reactive transport model. (1992). and processes have to be considered: pressure of liquid and gas) and of the meteorologi- cal data. atmospheric tem. and (2) steady. contents obtained by weighing the columns every culations we built a new code that coupled a code 24 h after a leaching event. Pg and PI are transport and chemical reactions. Meteorological processes (rainfall. Retention curves of the various materials (input • transport of gases both in the gas phase for CODE-BRIGHT) were measured in the (diffusion) and in the liquid phase (less laboratory by means of a transistor psychrometer important). These parameters were BRIGHT is the intrinsic permeability. material: persion and diffusion in the liquid phase. gas and heat as a of pyrite oxidation in the soil. solar radiation. Fig. evaporation and surface run-off.192 M. • chemical reactions such as aqueous speciation. Fig. and meteorological data (rainfall. the following function of the state variables (temperature. and hours of sun and wind sludge.

evaporation rises. The model was run using small time steps (at Figure 5 displays the results of the flux model most a few hours) in order to be able to see not of the column with the clayey mixture.4 clayey soil. Infiltration is defined as precipitation minus surface run-off. redox. Results of the flow and heat model for the sand-sludge mixture. bottom of the column is nearly 1. As a consequence of the high tion. The first 0.08m both in depth and in surface. One can clearly observe the rise of hydraulic saturation (up to 1) after each Reactive transport model leaching and rainfall event. In this case. rain showers.04 m contained only in the surface. can be assumed to follow either an equilibrium Furthermore. adsorption. precipitation saturation. effects of rainfall and leaching events follow the ations. Figure 4 shows the results of the model for the sandy mixture.e. and the remaining 0. Chemical reactions ation. the saturation is higher in the clayey column. temperatures at different depth follow practically the same evolution. in its turn.12m. causes a drop in temperature. complexa- leachate. This means that all Fig. Temperatures are given for one depth only because spatial variability is negligible. one can see the daily variation or a kinetic approach. . Triangles indicate leaching events. The only seasonal influence but also diurnal vari. 4. and dissolution of minerals). The models used a one-dimen. respectively. leaching events and effects of individual same pattern as for the sandy column. the contained a mixture of sludge and sandy or column of clay achieves a saturation of nearly 0. above all. is carried out using the global implicit or Saturation is generally lower and shows more direct substitution approach. on the other hand. The high evapor. temperature. the infiltration of the tion. appear to vary with depth. i. evaporation is commonly negative. sional domain of 0. The numerical solution in saturation and. dispersion and diffusion) and chemical rainfall was not enough to produce some reactions (acid-base reactions. does not clayey soil. while the saturation at the quartz sand only. Temperature. However. As flow entering the system is defined as positive. With the exception of RETRASO simulates transport processes (advec- sample 9 (Tables 2 and 3). cation exchange. OXIDATION OF SULPHIDES IN UNSATURATED SOIL 193 obtained when intrinsic permeabilities of 10 -17 temporal variation at the surface than at a deeper and of 10 -18 m2 were used for the sandy and the level.

temperatures and values O2(g) is expected to diffuse freely in the fluxes by means of CODE-BRIGHT and. evaporation is commonly negative. For these of first calculating pressures. equations are solved simultaneously by applying the saturation during the days after leaching. Hydraulic saturation distribution averaged in of the hydraulic saturation with time. the hydraulic saturation was close to 1 chemical reactions by means of RETRASO. 6. in the deeper half of the column. and partial pressures in the gas phase depending on consumption of O2(g) or CO2(g). SAALTINK ET AL. temperatures at different depth follow practically the same evolution. This simple method may be used when these parameters do not change significantly and/or small time steps are used. For Newton-Raphson. the hydraulic CODE-BRIGHT and RETRASO. which O2(g) did not easily diffuse. except for the immediate A simple time-lagged approach couples hours after a leaching event. gas phase. Temperatures are given for one depth only because spatial variability is negligible. Fig. but variable in space. Triangles indicate leaching events. The reactive transport model assumed that during the periods of outdoor exposure the flow of water could be neglected and that hydraulic saturation was constant in time.194 M. suggesting that Examples of such parameters are porosity. which is the case for our calculations. This consists saturation was always below 0. second. W. Results of the flow and heat model for the clay-sludge mixture. Infiltration is defined as precipitation minus surface run-off. of using the fluxes to calculate transport and however.e. .6. Hydraulic saturation was obtained by averaging the results from the flow and heat models (Fig. i. The time. This average is a simplification of the real process. The full line corresponds to the sandy soil and the assumed average value tends to underestimate dotted line to the clayey soil. In the case of the clayey mixture. 6). which involves the evolution Fig. depends on precipitation or dissolution. 5. As flow entering the system is defined as positive. the sandy mixture.

As no and oxidation by Fe3+: thermodynamic data are available for such a solid.007)(OH)3. These micro- total mixing. Minerals reacted with the pore water and modified the concentration of solutes. By assuming others) accelerate this reaction. From mass- reactions (Singer & Stumm 1970. Nordstrom & Alpers 1999): the results of a sequential extraction and analysis oxidation by oxygen: of the solid. The dissolution rate of pyrite. with an optimum growth at pH by: lower than 3. This holds promoted dolomite dissolution. The electron transfer from S2. the total liquid concentration of a organisms tolerate high metal concentrations and component in the leachate. Domenech et al. Dissolution con- for the longer periods of outdoor exposure. CL. the model included the Na. S/ its saturation and established. Likewise.K.016)O4]2(OH)6. the rate laws (see the next section). all of them with the same trace element proportion. The solution recovered after a rain event 5 days after leaching (sample 9 in Tables 2 and 3) showed It is believed that the oxidation of pyrite by O2 concentrations higher than expected from the (reaction Rl) is slower than the oxidation by Fe3+ chemical reaction rates. they also returned to the same initial situation.to SO 2 4 . Moreover. Domenech et al. minerals and gases required for describing the geochemical . sphalerite. There were identified by XRD and scanning electron Pyrite can be oxidized through two chemical microscopy with microanalysis. can be calculated are acidophiles.Pb0. the dependence of the pyrite dissolution column just before a leaching event. the reactive transport model Oxidation of Fe2+ by O2 has to precede the last assumed no flux during the periods between reaction: leaching. H. (2001b) leaching solution (distilled water equilibrated interpreted that the slowest steps of the oxidation with atmospheric CO2) and vw is the volume of process of the sulphides correspond to the the leaching water (vadded in Tables 2 and 3). calculated from kinetic (i.. cw is the concentration in the culations (Table 4). Micro- (piston model). Furthermore. This process pH of the leachate can be calculated from CL is equal for all the sulphides.018)Fe [(S0.e. the sludge contained small amount of other sulphides.993- Cu0. The same dependence has been cpi its total aqueous concentration as calculated assumed for the minor sulphides in the cal- by RETRASO. sumed protons and acted as a pH buffer of the because initial concentrations have a small soil pore water. For The acid conditions generated by the oxidation simplicity. Besides pyrite. through a standard chemical speciation. The most abundant minerals identified in the sand column after more than 400 days of Geochemical model functioning are gypsum and jarosite. Nicholson balance calculations in the leachates and from 1994.i is the mass of solute and determined by Domenech et al. OXIDATION OF SULPHIDES IN UNSATURATED SOIL 195 As stated above. than rather than a pure advective displacement therefore. At higher pH bacteria do not significantly accelerate the reaction over and above the abiotic rate. where vi the rate with pH and dissolved O2 concentration was volume associated to node i. (200 la) concluded that Cu coprecipitates with ferrihydrite in the initial stages of the experiment as (Fe0. dissolution and eva- influence but may give errors for the shorter poration caused the increase in ion concentra- periods (especially sample 9). Table 4 lists the aqueous species. mixing between the infiltrating organisms (Thiobacillus ferrooxidans amongst and the pore water also takes place. galena and chalcopyrite was where EvioSicp. reactions R2 plus R3).984As0. is the rate-determining step. In the case of the clayey soil. (2001b). K and H pure phases. More- EvioSi the volume of water in the pores of the over. tions and several new solid phases precipitated. This suggests slowest reaction of the last two is (R3) and. we assumed that after each leaching of pyrite promoted the dissolution of the silicates event the chemical composition of the pore water in the soil. As and Pb coprecipi- tates as trace elements in the jarosite. which could have an average stoichiometry of (Na.

(9)Busenberg & Plummer 1982. Pb0.1) 2 2+ 3+ 2() Fe /Fe (R3) r . (8)Swoboda-Colberg & Drever 1993.984As0. which accounts for a whose kinetic behaviour will be discussed below.2 s . to behave according to kinetic laws extracted homogeneous grain size and equal density for .016)O4]2(OH)6 Al(OH)3(am) H-jarosite (H.3soil) was roughly estimated minerals that dissolved. however.984Aso.9 X 10"8 [O2(aq)]0. from experimental studies. The silicate mineral( o m m 2 m .\ X 10*15 aH()~22) crcli 4 (Ocii ~ 1) * The equilibrium constants of the aqueous speciation. A1(SO4)2. Pyrite (Rl) r = 7. The rate laws for reaction (R2) and (R3) have been changed as explained in the text. (7) assumed equal to gibbsite. Table 4. (2) Nicholson 1994. Fe(OH)t2.7 X 10~ [O2(aq)]° a°H aGn (HGn .4 X 10~ [O2(aq)]° c$ <rSph (/2Sph . chemical saturation.1) 7 12 Hornblende r = 1. were assumed on a geometrical basis. Na+. Fe3+. SAALTINK ET AL. SiO2(aq). H2S(aq). K+ Ca2+. + Only for the model of clayey soil with sludge.018)Fe3[(So. Cu 2+ .4 ae'im °py ( A>y.". CO23.5 X 10~ 10 a&5+1. Fe(OH).1.1 ) 1 Pyrite (R2) r = 6. decrease in the rate as equilibrium is approached The gas species and minerals precipitating were (Lasaga 1984). H2AsO4- Other aqueous species (equilibrium) HSO 4 .1) 1 Chalcopyrite r . HCO 3 . (4) assumed equal to muscovite.0 X 10" n ^3VGlb (/2Gib .[(S0.2.4.( 4 .018)Fe3[(S0. Wieland & Stumm 1992.1) 6 Al(OH)3(am) r . SO2-4.(aq).1) 5 Albite r = (7. All the laws (except Reactions between aqueous species were that for R3) are multiplied by the reactive surface assumed to be in equilibrium except for the of the mineral and by a term dependent on oxidation of Fe2+ to Fe3+ (R3 above mentioned).8 X 10~ crHorn (/2Horn ~ 0 8 Clinochlore r .9 X 10"8 [O2(aq)]°49 0{V aCpy (/2Cpy . The reactive surface area of each also assumed in equilibrium.1 for reaction R3). Pb0. A1SO+4. Pb 2+t .l X 10~ 1 5 aH022) am 4 (^iii .5 X 10"13 a^()9+\. A spherical shape. (5) Schweda 1989. model with a reasonable degree of completeness. + References: (1) Domenech et al 2001b. HS . Al(OH)3(aq). C02(g) Precipitating minerals (equilibrium) gypsum Na-jarosite (Na.49 a^ opy ( O P y .3 s . gas and mineral dissolution are those of the EQ3 database (Wolery 1992).5 X 10~ 1 7 aH045) crFd (/2Fd .2 X 10" 14 «H° 3) <TP\ (/2P1 . 1996.018)Fe.6 X 10-9 aFe(II)-0. (3) Singer & Stumm 1970. Aqueous species gases and minerals used in the calculations * Aqueous primary species Fe2+. OH- Gases (equilibrium) 02(g). 2 x l O " 1 2 a& 38 +1.984As0. O 2 (aq). Zn 2+t . CO2(aq). (6) Chou & Wollast 1985.(1.oi6)O4]2(OH)6 Si02(am) jarosite (K.1) 1 Illite r = ( 4 . Mg2+. cr the reactive surface (m2 m-3soil) and n the saturation index. 2 x l O " 1 2 a&38+1. A13+.1. Mogollon el al. Kinetic laws: r is the rate ( m o l m . A1(OH)+. W.1 for minerals and m o l d m .0 X 10 a Fe(II) a02(ac()a^H+\ 3 X 10* ' flFe(II) a0^ 3 7 49 Sphalerite r .196 M.1) K-Feldspar r . FeSO4. Pb0.5 X 10"13 fl&09+l.016)O4]2(OH)6 ferrihydrite Dolomite+ (Fe 0.1) 1 7 49 l Galena r = 1. H+.993 Cu07)(OH) 3 Dissolving minerals (kinetics) Mineral Kinetic law Reference.0 X 10~ 10 a^5+2.

An average grain size of Although the model reproduces well the pH in 10-5 m was assumed for the sludge and the the last periods of the experiment. 4). . by the model. Nemati & Webb watering made the soil more humid (Fig. This disagreement is attributed This led to an overestimation of about 10% of to the fact that the model assumes a constant the value of the rate constants. precipitation of iron hydroxide is not predicted tures predicted by the heat model (Figs 4 and 5). Figure 7 shows the experimen- tal and calculated values for the column with sandy soil and sludge. where the most frequent (Lasaga 1984. 6). data for the first 3 months. It can be seen that the modelled SO4 and Fe concentrations rise and that where o is the porosity of the material and dm is pH drops due to the pyrite oxidation.3 m for the sandy soil and pH values are much lower than the experimental quartz sand. This temperature was higher than the experimental values because the slightly higher than the average of the tempera. the grain size (m).3 . according to the hydraulic saturation over time (Fig. As the pH is under- Calculations were performed at a constant estimated. This activation energies of the reactions involved. Results of the reactive transport model for the sandy soil mixture. 7. the modelled concentrations of Fe are temperature of 25°C. Nicholson 1994. the calculated clayey soil. Wet 1997). OXIDATION OF SULPHIDES IN UNSATURATED SOIL 197 all the components (sulphides or silicates) were Results and discussion assumed: Sandy columns. and 10 . Concentrations are in moldm . Full squares represent the predicted concentrations in the leachate. conditions hindered the diffusion of O2 into Fig. the saturation is not representative for the first stages average of which is between 60 and 80 kJ mol-1 of the experiment. Open squares represent the analytical values.

experimental pH jarosite (Domenech et al. This does not agree with the dynamic values of pure-jarosite terms. The calculations. precipitated form (jarosite). This indicates that literature are used. no experiment prevented the formation of the thermodynamic data are available. respectively. but. This indicates that Pb is not coprecipi. SC>4~. the amount of Na and constant of the oxidation of pyrite by Fe3+ K in solution did not increase due to the formation (reaction R2) by a factor of 5000 and to increase of jarosite. Similarly to Fe. thus controlling the amount of reduces the reactive surface (<j m ) and. no explanation was found concentrations (see below). This laws. In that expected to retain As. due to the exhaustion of the sulphides. As for Mg. amorphous A1(OH)3. and was surprising. Pb are orders of magnitude higher than the Figure 9 shows profiles of various modelled analyses. in accordance with the orders of magnitude. unfortunately. therefore. well by the model. When formed. the model cannot verify samples. consequently. the column and. Figure 8 shows the mis- experimental values that were below the detection match in the Fe2+/Fe3+ ratio if rate laws from the limit (c. No explanation for the high Mg contents at the Fe3+ was found both in the liquid and the early stages is given. the predicted concen. This is consistent with the the rate law constants for reaction (R2) and (R3) hypothesis that jarosite specifically retains As. conse- Ca in solution. reduced the oxidation although its decrease seems to be related to of pyrite. experimental values. 200la). reaction (R2) is approximately as reactive disagreement is attributed to the use of thermo. control the formation of calibrated rate law and the rate law from the jarosite. 7 that the experimental concentrations are overestimated.and K-bearing jarosites step (Singer & Stumm 1970). 2 X 10 -4 moldm -3 ). if O2 concentrations were lower. detected in microanalysis of this phase. Pb was not decreased more slowly. the release of Al. difference of the rate law for reaction (R2) is more tration of Na did not decrease with time. The low K concentrations predicted for literature for reaction (R3) can be explained by the early stages are due to the precipitation of the acceleration of this reaction by micro- jarosite favoured by the low pH of the organisms (Nordstrom & Alpers 1999). the low pH values described in the sulphide weathering environ- predicted by the model at the beginning of the ments (Scott 1987). Fe and Cu show a decrease in concentration of Ca is constant with time and the last part of the experiment. Moreover. Fe2+ by O2 (reaction R3) is the rate-determining firm the presence of Na. except at the early stages. The concentrations of Cu and As whether atmospheric conditions prevail in reality. As we did not carry for the high Zn concentrations in the early out O2 measurements. either adsorbed and/or case. Our model predicts atmospheric O2 solid phase. reduced and Similarly to Fe. caused by the low modelled Zones of Fe~+ near pyrite grains could exist along pH values. tation in the model. The predicted concentrations of transport of solutes between these two zones. Fe3+ and Fe2+ concentrations is the heterogeneity. even with O2 model is similar to the experimental values.198 M. However. An The low pH values of both model and alternative explanation could be the formation of experiment cause the dissolution of silicates an independent solid phase with a lower and. and especially K. and the calculated Al It can be seen from Fig. In order to reproduce Although the dissolution of silicates releases this in our model we had to reduce the rate law Na and K to the pore water. W. as reaction (R3). As a consequence. The variation of the different . the overestimation of As increased even more. In the case of K. The difference between the Na. This could be attributed to the dissolution and later precipi. the predicted that for oxidation of Fe2+ (reaction R3) by five concentrations decreased. the Fe(OH)3 is with zones of Fe3+ close to other grains. Mg. sludge column. predicted by the model are similar to the Nevertheless. Ca. Indeed. concentrations in equilibrium with the atmos- confirming that this metal is not retained in any phere. concentrations and reaction rates for the sand– tated as a trace element in the jarosite. which is not agrees with the experimental values. An issue that we have not concentration at the early stages can be explained studied but that may affect the relation between by the absence of amorphous Fe(OH)3 precipi. with these calibrated rate higher than the experimental values. Na solubility product. SAALTINK ET AL. but is not represented experimental values. The modelled concentration of quently. Plumbojarosite has been and K. 2001a). reaction (R2) would consume all available The concentration of Zn predicted by the Fe3+ leaving almost all Fe as Fe2+. reduces the dissolution rate according to Mg matches the order of magnitude of the the rate laws of Table 4. whereas generally accepted theory that the oxidation of microscopic observation and microanalysis con. the rate-determining step would be the coprecipitated. This is observed for the modelled values. According to this (Domenech et al. The modelled data for Zn. theory. would have to be. This tation of gypsum.

in general. OXIDATION OF SULPHIDES IN UNSATURATED SOIL 199 Fig. The only difference with the previous mixture do vary in depth (Fig. and the preci. 4 ( I I ) °>y (A>y ~ 1) and r3 = 1. parameters in depth is. O2 that is high enough to maintain the Fe2+ con- and CO2 can easily diffuse.3 X 1(T5 0Fe(°li) ^(III) Opy (A>y ~ 1) and O . the amount of Ca in jarosite is more important as a secondary mineral solution is controlled by the precipitation of than gypsum or ferrihydrite.3 X 10~9 fl^ flgjiii) ^py (A>y ~ 1) and r3 .0 X 1015 a Fe(II) aO2(aq)«oH + 1-3 X 10~6 « Fe(II) aO2(aq). because condition. The dissolution of dolomite maintains higher than the pH calculated for the pore water a high pH. Observe that reaction (Rl) is K and Na remain constant throughout the somewhat faster than reaction (R2). Observe that. Evolution of Fe^+ and Fe3+ concentrations for different rate laws. Figure 10 shows the results of Contrary to the sandy column.g. and blank experiments consisting of cannot diffuse upwards. Open squares represent the analytical values.6 X 1 0 . This means that all the amount of Fe3+ in solution. Furthermore.1. The reason of model is the hydraulic saturation used and the this variation is the distribution of the hydraulic addition of dolomite to the set of minerals saturation.9 a F e 0 . both CO2(g) is small. pressure of 0. (A) r2 = 6. the concen- the model for the column with clayey soil and trations and reaction rates for the clay-sludge sludge. Clayey columns. O2 cannot easily diffuse downwards no other sources (e. Moreover. (dolomite). Full squares represent the predicted concentrations in the leachate. 9). As no jarosite precipitates. emanating from dolomite dissolution. as discussed below. Under this wet that oxidation of sulphides takes place. which reaches values near 1 at the modelled. so that atmospheric centrations at very low values. within hydraulic saturation.0003 bar). Owing of Fe(OH)3 is also favoured by the rate of (R3) to the low hydraulic saturation of the model. there is a good agreement between measured and calculated concentrations except for pH.3 X lO^ 1 « Fe(II) aO2(aq). the conditions prevail within the entire column and dissolution of silicates is slow leading to low Al the concentration of O2(g) is high and that of concentrations. causes a lower oxidation of the dry half of the column. The precipitation Fe2+ released from pyrite oxidizes to Fe3+ and . 8. experiment. Here again. CO2(g) sulphate concentrations in the leachates. regardless of an extra source of calcium precipitating at all. the measured pH was sandy soil.0 X 102° aFe(U) «O2(aq)4H + L3 X 10 ~ 1 ^Fedl) fl02(aq). the last not gypsum. (C) r2 = 3. When the pore water is columns without sludge gave much lower pulled out of the column during leaching. (Table 1). This high pH also promotes the rate of (R2). Therefore. 10). together with the elevated inside the column (Fig. (Rl) and (R3) and the precipitation of ferrihy- pitation of amorphous Fe(OH)3. producing an increase of the sulphate concentrations are lower than in the the pH of water.0 X 1020 a Fe(II) aO2<aq)floH+1. the reaction rates of the oxidation of Fe2+ to Fe3+. which. 1. The high SO4 concentrations indicate lower half of the column. (B) r2 = 3. small. As is released to the atmosphere (having a partial expected from the higher hydraulic saturation. the reaction rate of pyrite compared to that observed in the column (Rl) is several orders of magnitude higher than of sand and sludge. gypsum) were identified and CO2. At neutral pH. In general. which controls drite are almost identical.

Du et al (1997) and Manning & port model was able to reproduce the exper- Goldberg (1997) confirmed that As could be imental results of the sandy soil column. of As. Zn and Pb released to the solution in the Therefore. Roussel et al. 9.200 M. SAALTINK ET AL. No explanation for Zn a proportion of 1 wt% sludge and a gravel . (2000). can be retained by modelling results described above. sandy column (Tables 2 and 3). and As and Cu onto used to predict the behaviour of the soil in two illite. Once the reactive trans- Morel (1990). Concentrations are in bar for gasses and in moldm" 3 for solutes. the immedi- adsorption onto ferrihydrite or clay minerals. Rates are in molmco^mnS^ 1 . who suggested that those Predictions. no experimental data are available on the Experimental results showed that the amounts solid phase after the experiment was completed. it was adsorbed onto Fe(OH)3(a). SO4 and Pb retention was found. W. Likewise. No ferrihydrite precipitates in the sandy column and no jarosite precipitates in the clayey column. precipitates as ferrihydrite. Distribution of the concentrations and rates versus depth after 100 days of the sandy column (continuous lines) and the clayey column (dotted lines). column. According to the experimental and elements. This fact agrees with Voight et al (1996). ate risk of water pollution could take place in Tentative calculations using data of Dzombak & sandy and gravel soils. (1999) and Savage et al. Fig. especially As. sufficiently to produce the observed more realistic scenarios: a sandy soil containing leachate concentration. we did not include these trace metals in clayey column were low compared to those of the the model for the clayey column. Cu. Contrary to the sandy precipitates as gypsum.

Open squares represent the analytical values. as expected for a lower lower than that for a sandy soil.3 .6 moll . Full squares represent the predicted concentrations in the leachate. Concentrations are in moldm"3. Concentrations are in mol dm . rather than silicate dissol- Fig. the concentrations of Zn pH of the solution. and the case of gravel. Evolution of pH and the main contaminants for various scenarios of sludge-soil proportions. the gravel was calculated represented in Fig. respectively). Dotted line: 10 wt% of sludge in a gravel (silicate reactive surface assumed 10-fold lower than the sandy soil). assuming a silicate reactive surface ten times In the first case. of 100 days without episodes of rain are In the second case. A period of 100 days with no rain was assumed. practically the same of silicate dissolution. 11.6. Nevertheless. . containing 10wt% sludge. The results for a period Zn and As. jarosite precipitation controlled the observed. the pH does not decrease is maintained close to 4 due to the buffering role indefinitely but falls to 1. Surprisingly. As less sulphide is as for the sandy soil. Full line: 10 wt% of sludge in a sandy soil. in sulphide proportion in the soil. 11. Spaced line: 1 wt% of sludge in a sandy soil. We will discuss the and As reached in this scenario are also far above predictive results for Zn and As. This is due to the fact that at dissolved. pH is higher.1 for pollutants of the system. OXIDATION OF SULPHIDES IN UNSATURATED SOIL 201 Fig. Results of the reactive transport model for the clayey soil mixture. the main most statutory limits (10 -5 and 10 . as described in the experiments. 10. lower Zn and As concentrations are such low pH.

Science of the Total Environment. GARCIA-SANCHEZ. 1999. A. early stages of the experiment. less As is coprecipitated and impeding the dissolution of silicate minerals and the concentrations in solution are higher than for precipitating amorphous Fe(OH)3 instead of the sandy soil (Fig. the weathered soils. LOPEZ-PAMO. with reactive transport at a mine drainage site. According to the predictions. X. As a result. kept the pH high. The model assumed the hydraulic EVK1-1999-00097 of the European Union and saturation to be constant in time. Zn is predicted not to depend Conclusions on silicates. It is interesting to note the contaminated waters range between 10 -4 and non-lineal behaviour of the system. On the other hand.5 to 65°C and 0 to 1 atm PCO • American Journal precipitated together with Fe and SO4 to jarosite. which. affected the rate of pyrite This work was supported by the contracts PIRAMID oxidation. W. ANTON-PACHECO. The low As and Pb content of the dissolves. E. E. E.. weathering and mobility of heavy metals in the we feel that it is more important that the model Guadiamar valley after the Aznalcollar spill. Zn is jarosite. Estudio de la infiltracion a traves de .6 and 3 x 1 0 . & PLUMMER..A.O. 2000. As less such as in the clayey column. 1999). nates. & which time hydraulic saturation was under. E. however. The Zn and As concentrations of high input of acidity. respectively (Manzano et al. for predictive purposes. south- simulation was good in the long term. QUEROL. except References during the first days of outdoor exposure. 10-4. of Science. The kinetics due to the oxidation of pyrite. ROBERTSON. west Spain. took place in both columns. A. M. Modelling of sulfide oxidation between the two columns. in its turn. Boletin sandy and clayey columns. This parameter controlled contrary to earlier descriptions in the literature. & consistent with that predicted from the sphalerite MEDINA. Thus. J. H.G. less mineral is formed. 282. C. tion. jarosite to form. but no pH values lower than 1. Although hydraulic saturation was high in the 41-55. 41. The presence of carbonates. J. Rather than the satura. GALARZA. the pH is not moll" 1 . D. 1982.. The high concentration of Zn in the leachates was CARRERA. & REJAS. providing Na and/or K which 1. as shown by the high ORTIZ. L.. silicate of dissolution of dolomite in CO2-H2O systems at minerals dissolved. We found. 11). pH dropped Journal of Contaminant Hydrology.N. On the other hand. J. subsequent to the experiment was not available. In the case of gravel. As these two solutes are limiting for tation with jarosite. and the subsequent Na and K release. Geologico y Minero. piritico remanente en la cuenca del no Guadiamar however.. during ALASTUEY. The low concentration of trace metals not retained in jarosite and the predicted in the leachate of the clayey column was not concentrations in gravel are similar to those for modelled because information on the solid phase the sandy soil. jarosite is formed.202 M. BLOWES. The results of the reactive transport model simulated the experimental data well. The amount of As in solution is model. predicted to vary with the extent of silicate dissolution.. & carbonate minerals that caused the differences FRIND.. but the the same order of magnitude predicted by the amount of pyrite. Cartografia del lodo SO4 concentrations of the leachates. it was the existence of dolomite or another BAIN. and between 10 . oxidation of pyrite GUMIEL. 45-78.. of the behaviour of the soil at a daily scale.G. G. JIMENEZ. BUSENBERG. Nevertheless. DE M I G U E L . J.W. the amount of O2 that could diffuse into the soil. 2001. hydraulic saturation.6 4. but only on the initial amount of The flow model permitted a detailed description sludge in the soil. This has led to REN2000-1003-C03-02/HID of the Spanish Govern- overestimation of the sulphide oxidation at the ment. 107-114. The contaminated ground water precipitation of jarosite was very important to below these soils show pH values between 2 and maintain the pH at a value of approximately 1. ALFAGEME. sand or gravel.. 112. 23-47. and less K and Na are released to the leachates could be explained by the coprecipi- pore water. G. Jarosite is the most behaviour of other soil types and other sludge abundant secondary mineral present in the contents.. LOPEZ. The The model also reviewed rate laws proposed for most important parameter extracted from the flow sulphide oxidation and found that the oxidation model and used in the reactive transport was the of pyrite by Fe3+ was not faster than by O2. W. A. Evolution of pyrite mud estimated. The results of the calculations are consistent The model could be used to predict the with the field observations. In the absence of carbo. GOMEZ.4 according to the calculations. J.. very different behaviour between the mediante imagenes Daedalus-1268 ATM. 242.D.9 have been when silicate dissolution is not able to buffer a measured. F. sensitive to the type of soil.C. SAALTINK ETAL ution. less silicate oxidation. such as in the sandy column. deep part of the clayey column.

M. AYORA. 78-WR-08. In: BALL. & DREVER. S. Mother the Total Environment.. 329.. 787-794. the Sixth International of Water Rock Interaction. E.. C. DOMENECH. K. Modelling the effect of chemical heterogeneity on ROUSSEL.I.. APARICIO. NORDSTROM. MANZANO. num. Australia. N. Tuolumne County. T. Geochem- Sludge weathering and mobility of contaminants in istry of acid mine waters. Tecnica.W. LOPEZ-SOLER.C. Aznalcollar mine (SW Spain). 1649-1656. & TURRERO. In: Proceeding of nology. 3557-3572. J.L. G.. D. PS. 1999. & AYORA. Z. E. GERKE. THOREZ.. 133-160.M. Resources Research. 1121-1123.B. 1989. & FRIND. Colorado. A.N. American anisms: Laboratory studies. QUEROL. Iron-sulfide oxidation mech- and dissolution mechanisms of albite. (eds) Short Course on Environ- DIMOS. M. ANTON-PACHECO. Reviews in DOMENECH. Internal Mineral Association of Canada. 89. J. P. NAVAR. E.. T. American Journal rated Hydraulic Conductivity with a New Closed- of Science. & STUMM... accepted. BLOWES. C. that eliminates mineral concentrations. BLOWES. New ICARD 2000. & TANG. 1997. 285.. M. A. DOMENECH. NORDSTROM.W. D. & BLOWES. Sciences. J. Ontario. power station. 87-112. ENRESA (Publication ferrooxidans. American 189-209. Journal of Geophysical Research. Water C. C. C. R. 296. SAALTINK.G.M. 1998. K. (CODE-BRIGHT) for the coupled analysis of Adsorption of copper at aqueous illite surfaces. Chemical kinetics of water-rock l 'Academic de Sciences Paris. WAYCHU- TOUTO.A. BARETTINO. New York. Column experiments and the full VAN GENUCHTEN. Atmospheric Environment. 1996. A kinetic model for Department of Civil Engineering Princeton Uni- biological oxidation of ferrous iron by Thiobacillus versity. C. 232-242. 478-486. 57-88. 51–69. D. 1985.. J. Canada. 1997. & WOLLAST. Numerical modelling of acid mine drainage To. 1996. J. A. Earth and Planetary interactions. C. 1991. MARTINEZ-PLEDEL. generation and subsequent reactive transport. Lelotte. JAMBOR.M. & PLANA. Environmental Science & Tech.H. E. S. Nickel Rim Mine site.. & CARRERA. Dison.VIC ROADS. SOLER. RETE. & FRIEND. GARRALON. TINGLE. A. R. & DE PABLO. 242. C. B. GUMIAL. Compete rendu de LASAGA. chemistry of Mineral Deposits. 807–813. Sudbury. GENS.729–765. A. Research Report. Mining and the Mineral dissolution rates in plot-scale field labora- Environment Aurentian Univesity Press. Proceedings of Sudbury '99. Kinetics of alkali feldspar water interface. 53. dissolution at low temperature.. Fifth International Conference of method for the direct determination of dissolved Acid Rock Drainage. 242. 33.A. Ontario. 1–7. Min eralogy of Complexation Modeling: Hydrous Ferric Oxide. & MOREL. 34. 1978. O. & ALONSO. Solid solution in. S. J. W. 02/92.C. G. SCHWEDA. saline media. J. Rotterdam. M. & LASAGA. 1998. Science of in pyrite and secondary weathering phases. tory experiments.. Chemical Geology.. U.W. SINGER.. Mineralogist. W. & Journal of Science. PC. K.S. 2001a. & MON- SAVAGE. to acid mine drainage generation and treatment . & FERNANDEZ.M. NEMATI. G... 145-154. Transistor Psychrometer.N. 166–185. Science. O. A.W. K. A. SUN. E. vol 6A. G. Minerals. H. accepted. mental Science & Technology. Journal of Hydrology. A 4009-4025. D.. Measurement of Soil Suction Using mental Geochemistry of Sulfide Mine-wastes. 1994. 1999.A. mathematical formulation for reactive transport LOPEZ-PAMO. 13. northwest Queensland. family. Engineering Computations.UA. Water loo. K. W. Science of Total Environment. 167.S. Calculating the Unsatu- dissolution rate law of gibbsite. J. CHOU. 30. FORSLING. W. J. 105. 609-612. & ALPERS.. X. J. P. MOLSON. 963-993. GANOR. Report IR/91-93. 2001b. 1999.. DZOMBAK.S. Numerical formulation for a simulator Du. OXIDATION OF SULPHIDES IN UNSATURATED SOIL 203 la cobertera de la F. & BIRD.L. J.. R. 31.O. Balkema.B..V. BRIL. H.K. Q. 1976... 209. (eds) The Environmental Geo- (SW Spain). Denver. OLIVELLA.E. C. O'DAY. 1970.C.. D. D. AYORA. U. Environ- MOGOLLON. 1993. J. Sudbury. DE PABLO. Practical Identification of Clay MAYER.. L.. Chemical Geology.. 2005-2011. 1219–1244. Surface MANTILLA.J. Arsenic speciation pyritic sludge spill and its effect on soils. 187.. & WEBB. A. J. Fe(III) concentration in acid mine waters.. Chemical Geology. 1984. Steady state kinetics NICHOLSON. & MCCLESKEY. BENNER. immobilization of As and Fe. D. atmospheric particulates around a large coal-fired John Wiley. In: PLUMLEE.C. ARRANZ. & GOLDBERG. G. 163-183..K. & soil affected by the Aznalcollar mining accident LOGSDON. C. CARRERA. In: SWOBODA-COLBERG. 1999. C. 1999. D. form Analytical Model Water Resource Program. 72. SCOTT. M. California. R. MAYER. Ontario. USA. 2000. Colorado. 178-187. Journal of Colloid and Interface Science. F. Belgium. H. 1987. MANNING. Society of Economic Geo- Oxidative dissolution of the pyritic sludge from logist. 1997. 15. F. 1999. Biotechnology and Bioengineering.M. and stability of Arsenic (III) at the clay mineral. 1996. B. AYORA.K. 1990. CUNNINGHAM.. Acidic mine The reactive transport model MIN3P: Application drainage: The determining step. J. acidification and solute leaching in overburden Evolution of sulphides-rich mine tailings and mine spills. M. Lode Gold District. and classification The impact of Aznalcollar mine tailing spill on of gossan-derived members of the alunite–jarosite groundwater. Littletan. Adsorption Applied Geochemistry. Economic Geology. . ORTIZ.A. ALASTUEY. The extent of the Aznalcollar NAS. 2000.

1994. 3149-3158. E. 1994b. A..E. D.204 M. Modeling of multi. & HENNET. 1992.J. Chemical fixation of arsenic in contaminated of kaolinite in acidic solutions at 25°C. PTACEK. E. 3339-3355. A. 633-643.W. & STUMM.W. Applied Geochemistry. 3137-3148. BLOWES.J. 56. WIELAND.O. et Cosmochimica Acta.C. in mine tailing impoundments: a numerical model. User's Guide and Model development and evaluation. D.J. FRIND.W. E. lations. 30. WUNDERLY. C. BLOWES. E. National Laboratory.. J.W.L. 1996. W. W.W. R. 11. Modeling of multi. Water Related Documentation. J. FRIND. C. M. Livermore. D.. Water Resources Research. 32.. D. 30. T. EQ3NR. Theoretical Manual. Dissolution kinetics 1996. BRANTLEY.L. WALTER. SAALTINK ET AL.. 2. UCRL-MA-110662PTIV. & component reactive transport in groundwater. FRIND. PTACEK. WALTER.L. VOIGHT..J. . Lawrence Livermore Resources Research. 3173-3187.J.O. A Computer Program for C. 1. 1992.. WOLERY.D. Geochimica soils. Water Resources Research. & MOLSON. S. geochemical aqueous speciation-solubility calcu- component reactive transport in groundwater. PTACEK. BLOWES. Sulfide mineral oxidation and Metal mobility in aquifers impacted by acidic mine subsequent reactive transport of oxidation products tailings discharge. California. & MOLSON.O...

UK (e-mail: k. followed by producing minerals.gandy@ncl. Department of Civil and Structural Engineering. The model also includes the oxidation of ferrous iron to ferric iron in an attempt to account for contaminant 'sinks'. or the evolution of their studies by Younger (1997) proposed that the ini- strength over the contaminating life time of a tial exponential decay was followed by the spoil heap. Early workers (Frost 1979. until these source minerals are depleted (Younger lations as additional information on temporal & Harbourne 1995). London. they cannot will continue for many decades. waters. This relates base accounting. Spoil heaps often contain significant quantities of required. . Long-term contamination by Jambor (2001). the ongoing seasonal oxidation of source mine- linity-producing minerals. & ROBINS. such as acid– establishment of asymptotic levels. as described water table fluctuates.j. are rals. which are with the possibility of a problem occurring rather a potential long-term source of contamination for than how severe the problem will be over the local water courses and groundwater.00 © The Geological Society of London 2002. and the results compared with results from a mathematical model for contaminant release and transport. Department of Civil Engineering. within the Oxic zone as the inadequate for a number of reasons. Laboratory unsaturated column experiments on colliery spoil from a well-documented site in County Durham have been set up to assess its contamination potential. copper). Existing techniques. and alka. 198. The random walk method. University of Newcastle. the modelled results compare favourably with the laboratory results and any discrepancies can be accounted for. 0305-8719/027$ 15.ac.ac. In general. in the case of pyrite. provide the basis required for predictive calcu. there is no suitable 1983) suggested that chronic pollution should method to predict either the longevity of last no longer than a few years. University of Sheffield. GANDY1 & KATY A.uk) 2 Groundwater Protection and Restoration Group.S. In reality. zinc. Glover (Younger 1997). Existing acid-base accounting techniques provide little information relevant to the prediction of field contaminant concentrations and time scales. ducts. later contaminant sources. if not millennia. Effective remediation requires an have been shown to follow an exponential dec- understanding of long-term discharge evolution line with time. such as calcite. such as pyrite. However. sulphate in determining both the timing and the duration and. particularly pyrite. which use primitive laboratory to the initial 'flushing' of existing oxidation pro- assays to estimate the balance between acidity.g. a form of particle tracking. iron. Newscastle upon Tyne NEI 7RU. Comparison of time scales for tran- weathering and dissolution of these sulphides sport and reactive processes provides a first step releases metals (e. released by Oxidative weathering of pyrite minerals. Sheffield S1 3JD. Laboratory and numerical modelling studies of iron release from a spoil heap in County Durham CATHERINE J. the relative rates of contaminant generation and attenuation must be evaluated then extrapolated to the physical scale and environmental conditions of real field sites. For robust prediction. 205 Geological Society. for example where ferric iron spontaneously precipitates as ferric Oxyhydroxide and is effectively removed from the transport process. such as pyrite. N.evans@sheffield. which are then released into the wider Iron concentrations from abandoned mines environment. Younger (1997) used the changes arising from differential reaction rates terms 'vestigial acidity' and 'juvenile acidity' to and transient hydrological fluxes are also describe these two components. (eds) 2002. which may last up to 40 years. or the evolution of their strength over the contaminating life time of the sites. Special Publications. Oxidative long term.uk) Abstract: At present there is no suitable method to predict either the longevity of contaminant sources within spoil heaps. EVANS2 1Hydrogeochemical Engineering Research & Outreach (HERO). Mine Water Hydrogeology and Geochemistry. they are more concerned sulphide minerals. P. 205-214. with vestigial From: YOUNGER. UK (e-mail: c. is used to transport iron and sulphate 'particles'. acidity into percolating of contamination. At present. Essentially.L.

GANDY & K. released by the oxidative success of a pilot-scale wetland. is thought to have nation potential of the spoil material. to assess the contami. ferric iron spontaneously precipitates as ferric This proved to be a singularly efficient con- oxyhydroxide and is effectively removed from the structed wetland (Jarvis & Younger 1999). 1). labo. County Durham. As a prelude to a full-site investigation. UK. Historical and geological overview • Mineralogical composition was character- The Morrison Busty spoil heap. which themselves produce a computationally fast model capable of overlie Carboniferous Coal Measures beds. J. Map showing location of Quaking Houses. 1. compost wetland was constructed in 1997 at the The oxidation of ferrous iron to ferric iron is head of the Stanley Burn to treat the pollution incorporated to provide an iron 'sink'. while the vestigial acidity is being depleted. The heap overlies techniques in object-oriented programming to glacial clay and drift deposits. with limited success. (1999) suggest that planning should be based on intensive treatment of discharges using active processes. 1). Fig. The Acidic drainage water emanating from the random walk method. ash. EVANS acidity referring to the initial acidity resulting from the flushing process. a form of particle tracking Morrison Busty spoil heap enters the Stanley used successfully by Prickett et al. near Quaking ized by X-ray diffraction (XRD). is used to transport iron ultimately the North Sea (Fig. such as lime dosing. Estimates of the lifetime of the contamination arising from spoil are necessary for an optimum choice of remediation options. which can buffer the pH and increase alkalinity. UK (Fig. applied to selected areas of the heap in an attempt • The material was dried at 60°C and then to encourage revegetation. through the spoil. simulating the large time scales required. followed by long-term passive treatment using constructed wetlands technology for the juvenile acidity. 15cm tall The pollution emitting from the heap. UK. substantially reducing oxygen supply to the reaction site and hence oxidation rate (Wood etal 1999). ratory unsaturated column experiments have been carried out at the University of Sheffield on colliery spoil from the Morrison Busty spoil heap covers an area of approximately 35 hectares and in County Durham. Following the and sulphate 'particles'. The model uses the latest shale. The spoil comprises colliery release and transport developed at the University waste consisting of grey and black weathered of Newcastle. where and prevent migration into other water courses. packed into 10cm diameter. A. Acidity generated by pyrite dissolution may be neutralized by carbonate dissolution. has • Surface mineralogy was investigated by received material from surrounding coal mines scanning electron microscope (SEM) with since 1922 (Kemp & Griffiths 1999). • Drill material from 6–7 m depth in the spoil heap was retrieved and stored with- out drying at <4°C. Wood et al.206 C. through the heap in 1986-1987 (Jarvis & dimensional mathematical model for contaminant Younger 1999). (1981) to Burn. The results increased after a deep road cutting was made were then compared to the results from a one. varies in height from 4 to 10 m. Effluent iron and sulphate concentrations are subsequently calculated and are used to estimate the contamination potential of Laboratory methods the spoil material. Houses in County Durham. coal and coal dust. which flows into the River Wear and model solute transport. Mining an energy dispersive element analysis faci- ceased in the 1970s and top soil has since been lity. and juvenile acidity referring to that generated during seasonal fluctuation of the water table. transport process. which uPVC tube with sintered glass discs at the . Dissolu- tion may also be physically inhibited by the armouring of pyrite mineral surfaces with iron hydroxides. a full-scale weathering of pyrite minerals.

kaolinite and dissolution and gradual exhaustion of small pyrite.Bmlmin. tained at 20 1°C. top and bottom to distribute influent and weathering products of pyrite and clay minerals effluent flow (Fig. from juvenile sources began to dominate. of acidity. Jarosite jarosite and iron oxyhydroxides located on the saturation indices calculated using PhreeqC surface of illitic particles. This pattern probably results from the 1991) revealed that the spoil was composed of. These secondary (Parkhurst & Appelo 1999) are between . illite. alu- • Effluent pH and redox conditions were minium and sulphate were high. 2). concentrations of iron X-Ray diffraction and subsequent semi-quanti. Iron and sulphate • Samples analysed for major cations by concentrations were again initially high. and the experiment restarted. concentrated nitric acid. Intermittent blocking of the column by iron oxyhydroxides forming in the effluent tubes occurred towards the middle of the run (results not shown). and 6) as the proportion of contaminants released • The laboratory temperature was main. were present in small quantities (< 2%) and were • Deionized water influent was pumped onto not detected by XRD. Otherwise. was not observed and is. Schematic diagram showing the set-up of laboratory unsaturated column experiments on colliery spoil. the most common the top sintered disc at a flow rate ranging mineral capable of buffering significant amounts between 0. polypropylene tubes and acidified using After this period the spoil was washed and dried. Calcite. although values recorded by probes housed in custom. and sulphate declined slowly over the course of tative analysis of traces with Siroquant (Taylor the run. not shown). Scanning electron microscopy enabled quantities of relatively insoluble secondary observation of additional small quantities of weathering products. column throughout the experiment.1 . These characteristics are typical of • The effluent was sampled daily into the vestigial flushing phase of Younger (1997). but inductively coupled plasma-atomic emis. not • Flow rates were low enough to ensure unsaturated hydrological conditions in the included in the following model. quartz. 2. rapidly approached asymptotic levels (see Figs 5 sion spectroscopy (ICP-AES). therefore.05 and O. in combination of juvenile pyrite weathering with order of abundance. IRON RELEASE FROM A SPOIL HEAP 207 Fig. Initially.10 . rapidly declined over a period of 3 weeks (results made Teflon flow cells. and partial blocking may account for the rise in iron concentrations observed between Laboratory results days 20 and 40. effluent concentrations of iron. such as jarosite.

The spoil is assumed to be homogeneous so that the diffusion coefficient in the air-filled pores is assumed to be constant with where Z is a normally distributed random time and depth. which is taken to be a 'characteristic length' of the spoil material. velocity. with the mechanical mixing component considered to be zero. Dispersion is considered to be a mechanical mixing and molecular diffusion. J. DL is the coefficient of longitudinal The coefficient of molecular diffusion (DD) is hydrodynamic dispersion. Measured pH was relatively stable between 2 and 3 during the whole run. Modelling methods Contaminant transport A one-dimensional particle tracking model has The direct result of the random walk model is a been developed to simulate contaminant trans. This is probably because the experimental material was much fresher than the bulk of the spoil exposed to weathering in situ. It random process and is represented by an is represented by: additional displacement to each particle without modifying the mass of the particle: where ct\ is the longitudinal dispersivity. GANDY & K. Solute major advantage of this method above others is components are represented by sets of moving that it can be implanted directly on top of any flow particles. Vis the average linear velocity and DD is the coefficient of molecular diffusion in the where V is the average linear groundwater spoil material. mean (Zheng & Bennett 1995). EVANS and 0. Iron. The implications of this sampling artefact for tip modelling are discussed below. / is time. V is the average groundwater velocity and DL the Weathering reactions coefficient of longitudinal dispersion. This value is significantly lower than that observed at the field site. The walk method (Prickett et al 1981). Rd is the retardation factor. Contaminant transport using the random walk method is characterized by a normal probability distribution. with each particle assigned a mass that model and it does not involve numerical represents the total mass of chemical constituent dispersion. based on the random converted to a concentration distribution. where D0 is the diffusion coefficient in air. Contaminant transport using the random walk method can be equated to a normal distribution. consistent with the dissolution of this min- eral. with a concentration Cs and a flux rate Q.208 C. Each particle is therefore moved over a phase is incorporated into the model by moving distance x to its position at time t according to each oxygen gas particle according to equation equation (2) (Kinzelbach 1986): (3). A. which is an of a normally distributed random variable and approximate expression of pore geometry. The advective Oxygen diffusion component of the advection–dispersion equation (equation 1) is solved by moving the particles at The longitudinal dispersion coefficient (DL) in the same rate as the average linear velocity of the equation (1) is a term that includes both groundwater. Fig. represented by: sion. and n dispersive movement the deviation from the is the air-filled porosity of spoil material. 3. C is the concentration of solute and CsQ2 is a sink function. to: . sulphate and oxygen particles are present within the model. The resulting path lengths are normally distributed Pyrite is oxidized by dissolved oxygen according with mean Vt and standard deviation 2DLt. r is where advective movement represents the mean the tortuosity of the spoil material. involved. x is the space dimen. particle (mass) distribution that can be easily port through mine spoil. 3. which fluctuates between 4 and 5. as shown in The molecular diffusion of oxygen in the gas Fig. variable with zero mean and unit variance.

Figure 4 8 s . and by pre-experimental treatment of material to largely preclude vestigial contamination sources produced during sample storage. This below: was achieved by arbitrarily multiplying the rate of ferrous iron oxidation by a factor of 103.g. Moses & Herman 1991. position in the heap. which is evenly distributed throughout the column. it is assumed there is no influx of water through the side bound- aries. and hence the rate describe only a small fraction of spoil heap– expression. Nordstrom 1982. and the only produced spontaneously precipitates as ferric source of iron and sulphate contamination is Oxyhydroxide according to: pyrite. corresponding oxygen concentration from pub- tion used in this study uses the square root lished data for 1 atm partial pressure oxygen in rate dependence on concentration of dissolved Wehrli (1990). Conceptual model where the square brackets indicate species In order to model the column of spoil material. the rate constant. Oxidation of organic matter is not con- sidered. The rate law used is shown bacterial catalysis occurring in the columns. multi- 1980. the that in the pH range < 2 the rate equation is given applicability of the assumptions was ensured by bv: the collection of experimental material from a deep. As the experiments from which oxygen proposed by McKibben & Barnes these data were derived were performed in (1986) which was also used by Stromberg & abiotic systems. McKibben & Barnes 1986. The The conversion of ferrous iron to ferric iron is material is also assumed to be homogeneous. The rate constants K0. Water enters only at the top of the column. can increase Contaminant sources reaction rates by a factor of up to 106 (Singer & It is assumed that the spoil was not previously Stumm 1970). where it is evenly distributed across the spoil surface. Although the assumptions above The rate constant. likely to be bacterially mediated. Singer & Stumm 1970. Upon hydrolysis. Sung & Morgan ciated with 'calibrating' a multiparameter. testing the basic model with as few parameters as died the effect of pH on the oxidation of ferrous possible far outweighed the disadvantages asso- iron by dissolved oxygen (e. Several authors have stu. Bacterial catalysis by. oxygen concentration and are calculated for the Williamson & Rimstidt 1994). IRON RELEASE FROM A SPOIL HEAP 209 Between a pH of 5 and 8 the rate equation is given by: The literature on pyrite oxidation by dis- solved oxygen has been extensively reviewed and numerous studies on reaction rates have been published (e. Wehrli 1990). The column is assumed to be unsaturated at all times and the flow rate to be constant. For a pH in the range 2-5 the rate equation is Pyrite is assumed to be oxidized by dissolved given by: oxygen only. pyrite surface area.1 and incorporates scaling for summarizes the conceptual model. They all agree process model. which is not involved in any other . for equation (7) depends on pH and mine environments. relatively unoxidized. it was felt that the benefits of oxygen concentration. In this particular case. Dissolved ferrous iron is oxidized by dissolved oxygen according to: Hydrological conditions Flow is steady state and vertical.g. and k2 depend on Moses et al 1987. the ferric iron exposed to weathering processes. The rate equa. for example. Ferrobacillus ferroox- idans or Thiobacillus ferrooxidans. it was necessary to account for Banwart (1994). Millero 1985. K1. has a value of several assumptions were made. activity and k.

* Parameters taken from Bronswijk et al. and therefore Density of pyrite (kgm~ 3 ) 5000* enters the column through the side boundaries Average diameter of pyrite crystals (um) 50* as well as at the top.74 between the spoil and the tube wall.125 reactions other than the oxidation of pyrite and Width of column (m) 0. Oxygen gas is assumed to diffuse Mass of pyrite (%) 3. The gas then diffuses through the air-filled pores. The order of magnitude agreement between predicted and observed iron and sulphate con- centrations is consistent with the interpretation of Fig. J. Model input parameters and reactions taking place within the cell. 1 ters the column both as a gas and dissolved in Longitudinal dispersivity (m) 0. Generalized conceptual model of a single finite- difference cell showing fluxes into and out of the cell. Although it is possible to smooth the results.2 the influent.1 oxidation by any remaining dissolved oxygen. Average moisture content 0.089 the oxidation of ferrous iron. Gaps in the laboratory data represent occasions when the column was saturated or blocked so that hydrological conditions were not steady state. and is one of the disadvantages of using the random walk method for contami- nant transport.002 Partial pressure of atmospheric oxygen (atm) 0. with laboratory results exhibiting a pseudo-exponential decrease in sulphate concen- tration over the simulation period. EVANS difference grid at each time step.210 C. 4. This is incorporated into Bacterial catalysis factor 103 the model by adding oxygen gas particles at atmospheric pressure to each cell of the finite. Time step (days) 1 while all ferric iron subsequently produced Effective porosity 0. Figure 5 shows the variation in observed and calculated iron concentrations for a 220 day simulation. it was not felt to be necessary in this study as the general trends were evident. Table 1. These were either measured in the laboratory or taken from published data. All ferrous iron Surface width of column (m) 0.089 produced by pyrite oxidation is available for Average Inflow (1 day -1 ) 0. (1993). within a single time step. GANDY & K. Parameter Value Number of nodes 5 Length of column (m) 0. The results show a similar pattern to that for iron release. . The slightly rugged appearance of the modelled results is due to the random element in the model design. Oxygen en. Contaminant sinks Ferric Oxyhydroxide (Fe(OH)3) is assumed to be the only sink for iron in the model. No sink occurs for sulphate. Results and discussion Table 1 gives the model input parameters used in the simulation. Figure 6 compares observed and predicted results for effluent sulphate concentrations from the col- umn. A. whereas mo- delled results reach steady-state conditions almost immediately.2 precipitates as ferric Oxyhydroxide.

laboratory results. . It is improbable component derived from vestigial weathering that the downward trend in laboratory results is of secondary minerals (not modelled). A reduction in reactive surface area be unlikely to affect iron and sulphate equally. however. Laboratory and modelled iron concentrations for a typical 220 day simulation. could preparation suggest that this is unlikely. as mass- interpretation is also supported by the agreement balance calculations show that only a small pro- between iron and sulphate ratios. i. could produce the same effect. with an additional results must also be considered. 6. 5. as large con. IRON RELEASE FROM A SPOIL HEAP 211 Fig. that the bulk of the Alternative explanations for the differences in dissolved contaminants are derived from pyrite behaviour between the predicted and observed weathering (modelled). It is also produce the observed pattern.e. unlikely that dissolution rates decreased as a * ig. This caused by the exhaustion of pyrite. extended preliminary testing and lack of sample which contains both iron and sulphate. whereas a small contribution from jarosite. portion of the total is accounted for in the tributions from an unmodelled process would effluent. Laboratory ana modelled sulphate concentrations tor a typical 22U day simulation.

from the spoil heap. The model shows an exponential decline iron to re-enter the cycle (equation 14): in concentration with time. the model underprediction of effluent iron and dation of pyrite is undertaken by dissolved sulphate concentrations. J. In the laboratory experiments. oxygen to produce ferrous iron (equation 12) Figure 7 shows a model prediction for the that is oxidized by oxygen to produce ferric long-term depletion of pyrite within the column iron (equation 13).1 after approximately 55 years. the pH within the column is signifi- droxide and is therefore available to react with cantly lower and the occurrence of pyrite pyrite. Iron overprediction by the model in the later stages of the experiment could indicate that the ochre - ous precipitation observed in the output tubes represented a significant unmodelled contami- nant sink. as this has only been observed to any significant degree in the presence of limestone or at neutral pH (Nicholson et al 1990. Puura et al 1999). In very low pH conditions dominant process. the overprediction by the laboratory results in the early part of the run would not be expected. The calculated half-life for iron is 38 years. if this was the case. A. 7. however. ferric iron does not precipitate as ferric oxyhy. with iron remaining in the effluent for up to 134 years. whereby initial oxi. Significant sulphate concentrations persist for Fig. However. EVANS result of pyrite armouring by iron oxide. GANDY & K. . The ferric iron subsequently in terms of effluent iron and sulphate concen- oxidizes pyrite. and oxidation of ferrous iron occurred to search is to predict the long-term contamination a significant extent within the effluent tube. and as the pH at a field scale The model does not incorporate pyrite oxida. but concentrations are predicted to fall below I m g l . It is likely that the ochreous material was deposited Oxidation by ferric iron was omitted from the during periods when the solution was essentially model because the overall objective of the re- static. oxidation of pyrite by tion by ferric iron. is on the order of 4-5. it assumes oxidation by dis. thereby producing more ferrous trations. dissolved oxygen is considered to be the solved oxygen only.212 C. This represents the cycle suggested by oxidation by ferric iron is a possible reason for Singer & Stumm (1970). Modelled long-term iron and sulphate concentrations.

1985.L.and neutralisation-potential values in acid rock concentrations are of the order of decades. CAMPBELL. & HERMAN. P.B.S.J. actions.J. 1983. Rate control by oxide coatings. & the Water Resource Systems Research Laboratory at the University of Newcastle. material. & GRIFFITHS. 1561-1571. L. J. A. Vol 10. depletion of pyrite within the column of spoil Geochimica et Cosmochimica Acta. H. Journal of the Chartered unsaturated column experiments. under their Environmental solution: 2. United Kingdom. E. BRONSWIJK. Mineralogical Society of Great Britain & Ireland. NERETNIEKS. J. suggesting that a small experimental column L.. Diagnostics Programme. 22. Water-Resources Investi- GROENENBERG. 1509-1520. 49.L. PARKHURST.O. 55.A.J.G. FANNING.K. output of the model indicate an oversimplifica- New York. Batch Reaction. Art. & YOUNGER. J. NORDSTROM. tion of the source–sink terms during model MCKIBBEN. E. C. Champaign. material from the Morrison Busty spoil heap in KEMP. & REARDON. Elsevier. FROST. D. & BATCHELDER.A. Geochimica et Cosmochimica Acta. 544-554.L. J. pyrite in low temperature acidic solutions: rate laws tude agreement between observed and predicted and surface textures. Anthropogenic Influences.S. in acid sulphate soils.S. 101-111. A method for predicting the contamination construction and performance of a full-scale potential of mine spoil is proposed that is tested compost wetland for mine-spoil drainage treatment by modelling weathering rates from laboratory at Quaking Houses. The relationship of mineralogy to Half-lives for both iron and sulphate effluent acid. Jon Carpenter. UK.E. Aqueous pyrite oxidation by material. Water Science & Technology. M. Banwart and HOSSER. Atmospheric oxidation of the pyritic waste rock in . 40. England. Restoration Group at the University of Sheffield. Oxidation of conceptualization. and S. The research project is funded by NERC (Grant 1990. & and the contamination potential of mine spoil MILLS. Contaminated Land and Waste Management. 1999. JARVIS. Madison. The calculated half-life for sulphate is 40 years. 471–482. T. The effect of ionic interactions on A model prediction is given for the long-term the oxidation of metals in natural waters. Evaluation of the rate of decrease in Illinois State Water Survey.G. for Selected Groundwater Quality Evaluations. HERMAN. (eds) Acid Sulphate Weathering. 1991. 1987. K. Special Publication. Approach to Water Pollution. agree with experimental Science and the Community: A Collaborative results on an order of magnitude scale. Journal of PUURA. Groundwater Modelling. In: KITTRICK.J. gations Report. C. I.S. Geochimica et Cosmochimica results represents a significant potential in Acta. However. NUGROHO. Hydrology.. dissolved oxygen and by ferric iron. County Durham. 1999. laboratory experiment and the steady-state KINZELBACH. Pyrite oxidation in carbonate-buffered reference GST/02/2060). VALSAMI-JONES. C. JAMBOR. R. D. R. 395-402.L. 1986. Geochimica et Cosmochimica the long-term severity of mine water pollution Acta.W. M. Users guide to PhreeqC (version 2): A Computer Program for Speciation. 1-103. 1993. Pyrite oxidation at polluting for several centuries. J.L. F. This emphasizes circumneutral pH. & LONNQUIST. H. erences between the dynamic response of the Charlbury. 15. MOSES. (eds) containing only 1 kg of spoil could. although they become overview of problems and control strategies in the fairly insignificant (< 1 mgl" 1 ) after 210 years. 1982. Design.. Diff. J. predictive capability for this type of system. Aqueous pyrite oxidation The authors would like to acknowledge the and the consequent formation of secondary iron contribution of P. NAYMIK. A 'Random Walk' Solute Transport Model Quality. 51. Conclusions 141-161. NICHOLSON.K. 313-318.D. W. CJ. P. I. 1979. K. R... Journal of Environmental 1981. 37-56.an for as long as 273 years. 50. the order of magni. mine shaft in County Durham. GILLHAM. and Northumbrian Water Geochimica et Cosmochimica Acta. Bulletin. ARIBAWA. Younger and other members of minerals. the iron content of water pumped from a flooded 65.. Quaking Houses. C. D. 2001. in fact. An Intro- duction with Sample Programs in Basic. Group Research Centre Ltd. I. A. A.G. L. Mine water pollution . J. & KIRSIMAE.R. One Dimensional References Transport and Inverse Geochemical Calculations. drainage.. MILLERO. In: COTTER-HOWELLS.A.. D. 547–5553. & BARNES..P.C. T. Environmental Mineralogy: Microbial Inter- remain polluting for several centuries. & APPELO. 59-70.O. Results for Institute of Water and Environmental Mangage- simulations of iron and sulphate release from ment. E. Modelling of oxygen transport and pyrite oxidation PRICKETT. US Geological Services.S... 13. London. Soil other members of the Groundwater Protection and Science of America. & RITSEMA. 99-4259. IRON RELEASE FROM A SPOIL HEAP 213 longer and are predicted to remain in the effluent GLOVER. NORDSTROM. 54. showing that the material could remain MOSES.B.V.

6. 1121-1123.L. Kinetics and product Long-term changes in the quality of polluted of ferrous ion oxygenation in aqueous systems. C. cos–benefit analysis of future less quantitative analysis of minerals using the full environmental management options for the aban- powder diffraction profile. 14. (ed) Aquatic ZHENG. N. M. Computer programs for standard. 457-466. To pump TAYLOR. WILLIAMSON.L.D. 311-336.L. 583-595. 1990. 194/195. S. & BENNETT. 167. 1–19. S. SUNG.J. Science. Kinetic YOUNGER.D. P. YOUNGER. J. 405-415. P. WOOD. W.C. Cosmochimica Acta. ground coal workings in Scotland. 1991. In: STUMM. B. & BANWART. 9.D. or not to pump'. . Journal of Engineering Geology. Chap- Natural Waters. Estonia. JJ. A. EVANS Maardu. 39. & ROBINS. K. kinetics and electrochemical rate-determining step SINGER. & HARBOURNE. 1: field study and modelling.214 C. 58.. New York. Applied Contaminant Chemical Kinetics. Powder Diffraction. minewater discharges from abandoned under- Environmental Science and Technology. 69-79. 1994. 1994. Redox reactions of metal ions at ment. mineral surfaces. P. 1999. Geochemistry. The longevity of minewater modelling of geochemical processes at the Aitik pollution: a basis for decision-making. Geochimica et drainage: the rate-determining step. & RIMSTIDT. J. man & Hall.S. & STUMM. 1970. P. London. The Environmental Geology. 1997. YOUNGER. W.C. John Wiley. Institution of Water and Environmental Manage - WEHRLI. Quarterly 561-568. 1980. 1995. Applied the Total Environment. Science of mining waste rock site in northern Sweden. B. GANDY & K. J.C. doned Durham coalfield. 32. STROMBERG. 9. Reaction Rates of Processes in Transport Modelling: Theory and Practice. G. Journal of the Chartered 2-9. 5443–5454. W. & MORGAN. Acidic mine of aqueous pyrite oxidation.

Bolivia 3 URS Dames & Moore (UK). Contaminant source characterization of the San Jose Mine. Special Publications. water supply of the Bolivian city of Oruro. and forming a planar surface at around . P. & ROBINS. with input data based on the observed water inflow distribution and calculated mine volumes.surface water courses carrying mine 3700m above sea level (asl) in the vicinity of water and leachate. Norway 5 Norges Geologiske Unders0kelse. Holymoorside. closure and flooding). This pathways' water supply is derived from wellfields in the Altiplano. WAYNE HOLDEN3. . 215-239. URS Dames & . Broadly. Empirical and modelled data suggest that no risk is (or will be) posed to Oruro's public water supply wellfields at Challapampa. at risk sources.00 © The Geological Society of London 2002. URS Dames & Moore Norge. UK 4 Seconded to URS Dames & Moore Norge. ARMANDO TORRICO2. N.S. and teams of scientists and engineers were allocated to characterize: This paper describes a selection of the findings of a project carried out by the environmental engi. A groundwater risk assessment has been carried out considering: (i) the current mine water pumping operation. Kirkebakken 8. JAVIER RODRIGUEZ2. . (ii) potential future mine flooding. Oruro. Bolivia DAVID BANKS1 4. OLA MAGNE SÆTHER4 5. EDUARDO AGUILAR2. (eds) 2002. Project background and objectives approach to the problem. Oruro. Chesterfield. 0305-8719/027$ 15.the wellfields of Khala Kaja and Challa- Moore Norge elected to take a nsk assessment pampita (collectively known as the From: YOUNGER. filled with glacial. • sources: neering consultancy. Transport of contaminants in groundwater in the Quaternary sedimentary aquifer complex surrounding the mine has been assessed by empirical data and hydraulic-geochemical modelling using MODFLOW and MPATH. ommendations for actions to mitigate the impacts * recipients. N-7491 Trondheim. 86 Holy moor Road. characterize the San Jose mine workings as a –groundwaterbodies already contami- nated potential source of contamination to the public by mining-related parameters. Avenida Santa Cruz #1344. and (iii) mine waste leachate. The project outputs were to include rec. of mining-related pollution. George's Road. ZELMAR ANDIA2. . London SW19 4DR. DAVID KOLLER2. St. UK 2 URS Dames & Moore (Bolivia). Wimbledon. La Paz.L. 5 St. Bolivia Abstract: The San Jose Mine is a mothballed Ag–Sn mine near Oruro on the Bolivian Altiplano. Mine water chemistry has been characterized by field analyses. Piso 1. N-4012 Stavanger. 215 Geological Society. Cochabamba. a high-level intramontane trough in the . Norway 6 Corporacion Minera de Bolivia. fluvial and lacustrine plex underlying and surrounding the city sediments and volcanics of Tertiary–Quaternary of Oruro. the objective of the project was to ~ associated mine waste areas. George's House (2nd Floor). London. Derbyshire S42 7DX. Continued pumping and discharge of mine water poses a potential risk to surface water recipients and private groundwater abstractions located alongside these. 198. age. CARLOS MENDEZ2. Mine Water Hydrogeology and Geochemistry. Mine flooding rates have been simulated using two models (MIFIM and MODFLOW). RENATO VENEROS6 & JORGE FLORES6 1Holymoor Consultancy. for the Bolivian State Mining Company the mine complex itself (considering the (COMIBOL) with funding from the Nordic possible implications of future mine Development Fund and Bolivian Government.the Quaternary Altiplano aquifer corn- Andes.

Fig. Site plan showing the city of Oruro.216 D. BANKS ETAL. mine water channel. River Tagarete and selected monitoring wells. . San Jose Mine. 1.

Oruro. 1986-1994 +24. daily temp. Shallow-level mining.340 and . Three shafts allow access to the deep mine This paper focuses on characterization (in interior from the surface (see Fig. 24 h precipitation (airport) 34mm • pumping ceases at San Jose Mine.the aquifer complex as a potential future groundwater resource.30 m). 2). exploitation commenced in 1948.1 (average 8. papers. beneath the city of relatively well mapped. and its exact extent is aquifers. Vinto Meteorological Stations (after SGAB-COMIBOL . the focus of the world-renowned There is remarkably little seasonal fluctuation. The paper workings. Climatic data monthly pumping rates in the Santa Rita Shaft for the area are given in Table 1. Mine Mean actual annual lake 1875mm year. CONTAMINANT SOURCE CHARACTERIZATION. in the area now average of 3. a high-level the Auxiliar Shaft to level –340m. pampita wellfields. SAN JOSE MINE 217 Challapampa wellfields).other public and private groundwater 1995a). The Santa Rita and San Jose shafts allow and will be dealt with in more detail in future access to the — 340 m level from the eastern side. primarily as a tin and silver mine. The first was around 81s ] (Fig. production migrate towards the Khala Kaja-Challa. will also illustrate the range of parameters that provides additional access to the –460m level need to be quantified to generate the source term from the — 340 m level. The objective of this paper is demonstrate Galleries at the . The total water make of San Jose Mine is located adjacent to Oruro city. rising approximately 300m the Santa Rita Mine entrance (level . . 1986–1994 . above the Altiplano around Oruro. bulk of the precipitation falls from October to March . which may reach as far as Oruro. the mine is thus collected at the — 340 m level. The abstractions. Average temperature +10.17. Modern tin and zinc-silver-lead • there may thus be a head differential bet. ranged from 5. supplying Table 1.1 I s .1 water rises until it overflows at the lowest evaporation (Poopo) available mine entrance. The mine was nationalized in 1952 (along with • mine water may thus enter the Quaternary other major Bolivian mines) and passed into the aquifer system via natural fracture systems control of the Corporation Minera de Bolivia or via unmapped mine openings and (COMIBOL).an introduction level — 340 m is lifted from level — 460 m through Oruro lies on the Bolivian Altiplano. of Socavon. 1). Oruro carnival. An interior shaft. in a series of low hills (referred to hereafter as the before being pumped up the Santa Rita Shaft to 'Oruro massif). entering the mine above — 340 m is allowed to lleras Occidental and Oriental of the Andes. The Itos terms of flows and chemical quality) of source shaft extends from the surface down to below the terms for the risk assessment. however. Other aspects of — 460 m level on the western side of the Oruro the risk assessment are briefly touched on here. For economic reasons. has a history of through groundwater in the Quaternary several centuries. Water entering the mine below The San Jose Mine . Most of the deep mining has taken place in the • in addition.71s. Water plain (c. Santa Rita and San Jose and hydrochemical model of the mine. not well documented. Of this total.1 was pumped up from levels occupied by the Plaza of the Church of the Virgin deeper than –340m (via the Auxiliar Shaft). an mine was that of San Miguel. The mine's From January 1996 to August 1999. 3). with evaporation from SCAB (1996b). may migrate towards the wellfields both above and below level 0.5°C The postulated risk scenario is defined as Min. the average 'level 0' is defined at 3800m asl. The drain down to — 340 m. for a risk assessment. groundwater contaminated by latter half of the twentieth century and hence is mine waste run-off. massif. The San Jose Mine has a history of over 400 although for the vast majority of months the rate years. the head in the Quaternary aquifer system.1 ) . nightly temp. Annual average relative humidity 43% • the Khala Kaja-Challapampita wellfields create an area of depressed piezometric heads in the Quaternary aquifer complex. was halted in July 1992. Since closure in 1992.5°C Max. when control of ween the water level in the flooded mine and the mine passed to the 'Banco Minero de Bolivia'.91s. Climatic data derived from Oruro Airport and Oruro with its drinking water (see Fig.460 m levels (Flores the construction of a conceptual hydrogeological 1998) connect the Itos. 3700 m asl) situated between the Cordi.7 to 13.3°C follows: Average annual precipitation 345mm Max. the workings have been kept dry by pumped dewatering to the — 460 m level (3340 m asl). the Auxiliar Shaft.

2. Schematic cross-section of the San Jose Mine. .Fig.

1996–1999. The pumping increase in August 1999 was related to the need to fully dewater the mine system to allow access for investigators. Pumping rates from the San Jose Mine. ((Cu. tin content diminishes surrounded by the dominantly Quaternary sedi. Ag and The bedrock Sn were considered commercially important. Turneaure (1960a. mentary complex of the Altiplano. and silver content increases. Sn and Sb as the principal metals at San Jose. into the deposits typically exhibit a mineral assemblage Rio Tagarete system (Fig. Ludington et al. At San Jose. Flores 1998). geology and glacial (Pleistocene) or possibly interbedded mineralogy are described in great detail by volcanic deposits (Dames & Moore 1967). an early tin stage and a late silver The San Jose Mine is located in an 'island' of sulphosalt stage (SGAB-COMIBOL 1995a). gold (Au). argentite tems and sediment quality (SGAB-COMIBOL (Ag2S). two mineralization stages Geological and hydrogeological setting occurred. CONTAMINANT SOURCE CHARACTERIZATION. marcasite (both FeS2) and pyrrhotite (FeS). Pescod & Younger 1999).) The (untreated) mine water discharges from the polymetallic vein (BPV) type deposit. Ag. (ZnS).As)4S13). i. and many other sulphide and complex sulphosalt minerals. It is classified as a Bolivian groundwater resources. cassiterite (SnO2). BPV Santa Rita adit. 3. chalcopyrite (CuFeS2). tetrahedrite 1995a. where consider. 16 Ma old. of which Pb. exposing the central conduit. intruding Palaeozoic metasedimentary bedrock. With increasing depth. and also including sphalerite into the shallow Lake Uru-Uru. composed to buried bedrock ridges. via a mine water channel. The San Jose deposit The Quaternary sedimentary complex corresponds to a Tertiary (c. native able damage has been wrought on lake ecosys. The Rio Tagarete is dominated by pyrite. The Tertiary igneous complex hosting the San Jose Mine is situated in black slates of the Silurian Uncia Formation. Avila of the Altiplano Salinas 1990) igneous volcanic-intrusive caldera The flat plains of the Altiplano are underlain by complex that has been eroded and partially Quaternary sediments of varying thickness (due removed. arsenopyrite (FeAsS). b. galena (PbS). non-equilibrium pumping. The sediments are typically fluvio-lacustrine (derived from erosion The San Jose ore of adjacent mountains and hill ridges). 1971) and Kelly & The Quaternary complex contains valuable Turneaure (1970). Tin Tertiary igneous rocks (the 'Oruro massif) mineralization predominates near the surface. ('Santa Rita' represents the total amount of water pumped from the mine. 1). (1992) regard Pb. 'Auxiliar' represents the water ingress derived from below level –340m.Fe)12(Sb. SAN JOSE MINE 219 Fig. Zn. alluvial troughs and of intrusive rocks (Long 1992.e. which are believed to . with some The San Jose deposit's genesis. shallow bedrock platforms).

4. The map is based on interpretation of geophysical profiles and drilling logs. . BANKS ETAL Fig. Contour map (m asl) of the base of the Quaternary Altiplano aquifer complex.220 D.

possibly indicating that any physical feature. Abstractions of importance for the city of Oruro include the following. It appears that there is a buried channel more than Fig. Test pumping of boreholes • currently. the divide would probably migrate south. complex in the vicinity of Oruro. based on sampling the combination of great depth and considerable and analysis of 1999–2000. Direct recharge through the surface of the Altiplano is believed to be very limited due to high evaporation and the low permeability of the surficial sediments. Groundwater salinity (Fig. such as borehole PP-117 at the airport (up to 8 I s . the levels for salinity due to the extreme lateral and main direction of groundwater flow is vertical variability). • the mine water discharge channel enters the • a small amount of groundwater is attracted Tagarete very close to the groundwater to the wellfields from the south. Quaternary aquifer sediments typically varies San Miguel. structure that is a very effective 'sink' for abstracting groundwater. the contaminated groundwater suggests that the hydraulic conductivity of the from mine wastes at San Jose. • the well field at Challapampita. There is a tendency for fresher waters • the groundwater divide is not caused by in the Khala Kaja area. A contour map of the base of the Quaternary aquifer (Fig. vertically. pumping water from a complex of semi-confined and con- fined aquifers. 5.. the divide piezometric surface in the aquifer complex (it has would move north. Santa Rita. Map showing concentrations of chloride in groundwaters from the Quaternary Altiplano aquifer 150m deep in the Khala Kaja area. pumping water from two boreholes at a rate of some 401s-1. lies to the south of the between 1 and c. caused by human abstraction. If contaminated water from these . The map indicates head southwards towards Lake Uru-Uru. In this case. the wellfields at Challapampita or Khala • water abstracted at these wellfields is Kaja. • smaller production wells. gradients to be low.1 on average) and those operated by private concerns and industries. north and NE. on the basis of flow dynamics. from around seven bore- holes. pathway for recharge waters from the foot of the Should abstraction increase in the future. largely derived from the NW. 6) has also been prepared of the Should abstraction decrease. not been possible to correct measured water • to the south of the groundwater divide. Symbol sizes are based on thickness of permeable strata produces a the boxplot (inset). The groundwater level structed on the basis of the contours map indicates that: suggest that contamination would migrate east from Oruro and then south. head around the Challapampita-Khala there is thus no risk from the mine wastes to Kaja wellfields. • there is an area of depressed groundwater • currently. CONTAMINANT SOURCE CHARACTERIZATION. Cordillera Oriental. etc. SAN JOSE MINE 221 be mostly recharged by run-off through coarse alluvial fan-like material at the foot of the mountain ranges of the Cordillera Oriental. 5) in the aquifer an ENE-WSW-oriented groundwater complex varies considerably. Groundwater flow lines con- Moore Norge 2000b-d). 4) complex was prepared on the basis of geophysical and drilling investigations.1 .1 (see Cortez & Torrez 1996). It is a dynamic the Khala Kaja buried channel is a direct flow feature. A map (Fig. creating divide. at a rate of some 1501s. which are defined as risk receptors in the risk assessment approach: • the wellfield at Khala Kaja. both laterally and divide to the north of San Jose. 50mday -1 (URS Dames & divide. Concentrations are in m g l .

222 D. Groundwater level contour map for the Quaternary Altiplano aquifer complex around Oruro. The map is based on empirical water levels measured in wells and boreholes in December 1999. Contours are in m asl. 6. BANKS ETAL Fig. . Note the very shallow gradients and the ground water divide north of Oruro.

7. This does not necessarily imply a risk. via fractures. The production approach. Fig. implying that contamination from the mine may then migrate northwards. Mine production • inflow of saline formation water from deep records for 1960–1992 were then examined to levels in the Tertiary crystalline massif. however. and calculate the total mineral production of the mine inward hydraulic gradients thus prevent direct (4 389 825t). systems in hard rock or (ultimately) from the surrounding Quaternary aquifers. These models require the following input data: Waste production approach. the MIFIM (Banks 2001) and around 2000000m3 from the deeper levels. Mine water hydraulics At present. Given that intensive is closed down. Cumulative distribution diagram of mine void space (as void below a given level) in the San Jose Mine. This approach • inflow of groundwater from regional yielded a volume of 2196000m3 void below aquifer storage in saturated fracture level 0. CONTAMINANT SOURCE CHARACTERIZATION. Void distribution in the mine Water influxes to the mine The volume of open space in the mine was There are several possible sources for the water estimated by three approaches. Again. 1863 000 m*. but most probably galleries and workings in veins was estimated through historic mine openings in the on the basis of their physical dimensions. • the level of the most likely overflow All three methods of mine void calculation horizon. MODFLOW (McDonald & Harbaugh 1988) models were employed. To answer extrapolate production backwards in time to these questions. measured from mine plans. necessary to consider a future the total volume of material removed from the scenario where the expensive pumping operation mine was 1463 275 m3. a volume of 2 662 000 m3 mining of the mine host rock and the recharge rate waste corresponds to an original rock volume of (MODFLOW). and tailings around the San Jose massif as • the locations andflowsof major influxes of 2662000m 3 . mining waste of 30%. this does not necessarily imply a risk. Assuming a porosity for the mine water to the mine (MIFIM). . being pumped from the mine: • direct infiltration of precipitation from The geometric approach. It gravity of 31 m-3 for the extracted rock material. The volume of above. as there is evidence (URS Dames & Moore Norge 2000d) that the aquifer system attenuates mining-related contaminants extremely effectively. there is a possibility that it may flow slowly northwards toward the wellfields. and an original rock • an estimate of bulk hydraulic conductivity porosity of 0%. the ground- water divide may move south. — 80-m level (potential the mine would subsequently overflow (if at all) overflow level . • if abstraction is increased in the future from the Challapampa wellfields.see below) is believed to have at the surface and the rate at which it might leak commenced in the 1940s or 1950s. distributed according to Fig. SCAB (1996c) • the distribution and total volume of void have estimated the total volumes of waste rock spaces in the mine. is. Based on an assumed specific leakage of mine water to regional groundwater. SAN JOSE MINE 223 watercourses infiltrates the aquifer (and there is evidence that it does: URS Dames & Moore Norge 2000c). suggest that a floodable void space of some 2 000000–2 500000m3 is realistic. based on the geometric approach (see text). 7. as surface. and to estimate the rate at which production below the c. the mine is being dewatered. one can to regional groundwater systems.

1. however. BANKS ETAL. 11 s"' unaccounted for water from level deeper levels (–340m and below). If we assume that Modelling of mine flooding all the rainfall (364 mm year. 2) at the — 80-m • there is little soil cover and vegetation. the amount of run-in is Survey MODFLOW model (McDonald & less. . This corresponds well to Darcy's equations for porous media. this totals an amount of 75 440 m3 flooding.1 : SCAB. Conventional models for calculating indicate the most probable pathway for water recharge to the San Jose Mine from precipitation overflow from the mine in the case of pump are of no value for several reasons: switch-off to be via the Santa Rita Shaft through the so-called Socavon adit (Fig. slope is high and the run-off coefficient is thus The MIFIM model focuses on the geometry of likely to be relatively high. and the US Geological lost to evapotranspiration. The MIFIM model of Banks (2001) only a proportion of the water falling on the has been codified for the Oruro mine in an catchments runs into the openings.224 D.51s" 1 unaccounted water inflow systems (D. e. 19966) falling on the openings and their catchments runs In order to simulate possible future mine into the mine. 1. • c. two different models have been year"1 or 2. Flow gaugings have of scenarios.). from the periphery of the mine system at the • c. etc. 4 1 s . Given the uncertainty surrounding the origin of mine inflow.21 s from an exploration borehole in although salinity and isotopic evidence suggest the Gartner Vein at the . while 6 .1 ) . the mountainside. Church of the Virgin del Socavon. The results indicate that the inflowing sources (Fj). and the topographic two models use differing basic philosophies. the A field reconnaissance of exposed mine open. The soil thickness and vegetation cover are Harbaugh 1988) has also been employed. are dewatered and any potentially transmissive fractures at higher levels are now dry. Given the unreliable or non-existent mapping of many of the older mine workings. and assumes discrete water inflows mine via run-in to mine openings and a small (fractures. • c. Anecdotal evidence exists. Possible water overflows from the mine Ingress of surficial waters (head . Undoubtedly. . 2. modelling has employed a range Water inflows to the mine. level. Salinas.81s"1 inflow to the -340 level near see below.independent Field surveys and examination of mine plans inflows). all three components exist. possibility of mine water overflowing elsewhere ings allowed calculation of open areas and the than at the Socavon adit cannot be excluded with topographic 'catchments' of these openings on certainty. The total area draining towards mine openings and worked vein outcrops is estimated as some 207 250m2.51 s-1 inflow from Vein D at the . we assume that employed. a major component of deep-seated saline water . . events or works on surface water drainage • c. however. but does estimates made from hydrochemical evidence not realistically attempt to simulate the mine (see below). Hydrological the mine void and uses simple hydraulic estimates suggest that the total recharge to the algorithms. COMIBOL pers.). This currently contains a mining museum • there is a very rapid surface run-off and has a sulphurous-smelling entrance from the response to rainfall. The final and • old workings of exposed veins have left lowest adit exit is (ironically) within the grounds wide openings on the hillside that provide of the Geological Survey of Bolivia's (SERGE- the potential for direct inflows of rain OMIN) offices. for ingress of surficial waters. exploration boreholes. temporary • c. below from below level — 340 (given that Observations suggest that the majority of the total pumped amount from below significant water inflows to the San Jose Mine are — 340m in the Auxiliar shaft is c. If. comm. 1. geometry. the rest being EXCEL® spreadsheet.g. and of the order of 1 -21 s"].7 I s . This may — 340 or above (given that the total mine reflect the fact that the upper levels of the mine inflow is 81s" 1 from pumping data). F n ) . water into the mine.340 level. A best guess assumes that 1 -21 s"' been carried out of the major water inflows in the is derived from surficial head-independent mine. MOD- amount of direct ground infiltration is likely to be FLOW uses the porous medium assumption. the Santa Rita Shaft.420 changes in mine water flows related to rainfall level.4 Is" 1 . The low on the Oruro massif.1 is derived from water is derived from relatively few features: head-dependent groundwater inflows ( F { .

56 3. regional groundwater are estimated as less • . CONTAMINANT SOURCE CHARACTERIZATION.29 0 0.51s"1. flooding of the mine to level — 80 occurs This results in the following distribution of the within between 5490 and 7520 days (15. builds strongly on the GRAM (Groundwater The model has thus been run with the flow Rebound in Abandoned Mine-workings) model distributions shown in Table 2.1 of mine water are 1 • the 'missing' 11s" from the —340m derived from head-independent surficial level or above was assigned to the — 340 water sources.420 3380 3. respectively). A volume input to MIFIM was based on the crucial factor in determining rate of filling the cumulative volume distribution shown in and final discharge flux. above them. levels below –340.51s.63 2. 0.5 I s .1 .71 8 1. It will to account for variations in mine volume with be noted that the head-independent inflow F\ is depth. SAN JOSE MINE 225 which decrease as water head in the mine rises • -420m level: 3.e.1s. In many ways.51 s"! to level .00 2.25 0 0.5 and 1. it is less sophisticated than inflows F\ -Fn are all assumed to be equal to the GRAM (no Bernoulli-flow along roadways and head in the regional Altiplano groundwater no interchange of water between mine ponds). The external of Sherwood (1993) and Sherwood & Younger driving heads H\-Hn for the head-dependent (1997).51s.61s. The MIFIM concept.50 3.75 F2 .0 1 81s" inflow.44 0.1 .44 0. than 0. where . A worst case assumes 4 1 s . i.50 0.460m level: 0.1 .51 0. and (ii) it is able to use a reference table Six runs of the model were performed.5 25 50 100 25 F.340m level: 4. head-dependent flow and surficial • of the 1. level. Fig.38 0.25 F3 . The time step However: (i) MIFIM is able to consider non. As regards water inflows.employed is between 20 and 50 days.380m level: 0.1 ) .38 0.49 1. 0.50 0 2.340 3460 4.50 0. Of these figures.00 0 2.47 Discharge to groundwater 0 0.00 2.38 F4 .1 . as