Abstract Coal has hitherto been seen as a potential fall-back resource as hydrocarbons become depleted. However, amidst anxieties over peak oil and gas, some recent studies have painted a picture almost as gloomy about the longer term prospects for coal. Such evaluations are misleading, as they identify as reserves only that coal accessible at reasonable cost by means of conventional mining, which are increasingly modest compared with deeper-seated coal reserves amenable to underground coal gasification (UCG) using directionally drilled boreholes from the surface. Significantly for the petroleum industry, the technological requirements for UCG are far more akin to those of oil and gas production than they are to those of deep mining. A number of projects around the world are revealing the feasibility of UCG. We highlight preliminary findings from a recent investigation of the potential for UCG in NE England, which has the longest history of conventional coal mining at industrial scale anywhere in the world. Despite this history, fully 75% of the coal resources in NE England remain in place. A significant proportion of these is likely to move to the ‘reserves’ register as underground gasification technology begins to be deployed. A particular attraction of UCG lies in its suitability for coupling to CCS: we can use our long-standing knowledge of the response of incumbent strata to longwall coal mining to predict substantial increases in permeability in and immediately above the voids created by gasification. As these engineered zones of high permeability will already be connected to surface power plants by the wells and pipelines used to produce synthesis gas during gasification, they represent ideal prospects for permanent sequestration of a large proportion of the carbon dioxide arising. Stored CO 2 will be kept in place by cap rocks higher up in the sequence.

Abstract Although mining is no longer a key industry in the UK, the international mining industry continues to expand. One of the principal legacies of past mining in Britain is water pollution emanating from abandoned mine voids and waste rock depositories. This has necessitated many expensive technical evaluations and remedial programmes in recent years, from which important lessons may be drawn for the still-growing mining industry overseas. 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. As this phase is also the longest in the overall life cycle of any mine, it should be given appropriate consideration from the outset. 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. 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. The uncertainties related to predictive modelling of mine water arisings are considerable, whilst those of predicting mine water quality are even greater. Numerous pointers towards robust mine water management strategies are identified, and a call for ‘defensive mine planning’ is made, 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. With such measures in place, and enhanced monitoring data to hand, the conjunctive application of physical and geochemical evaluations will eventually provide much-needed predictive tools to inform site management decisions in the future.

Abstract The hydrogeological effects of longwall mines are vertically zoned. The heavily fractured strata immediately above the mine dewater, but they are typically overlain by a zone of low permeability that prevents shallower aquifers from draining to the mine. However, shallow bedrock aquifers experience head changes caused by fracturing during subsidence. New fracture void space takes up water, causing large head drops especially in confined aquifers. Increased fracture permeability affects heads because upper aquifers in high relief areas lose water through fractured aquitards to lower aquifers, and because the higher permeabilities lower hydraulic gradients and up-gradient heads, and increase down-gradient discharge. In addition, a secondary drawdown spreads out laterally through transmissive aquifers from the potentiometric low in the subsiding zone. After undermining, water levels may recover due to closure of fractures and to recharge flowing back into the affected area. Studies at two active longwall mines in Pennsylvanian coal measures in Illinois support the conceptual model, with variations. Unconsolidated, unconfined aquifers were not significantly affected by mining. At one site, heads in a moderately transmissive sandstone declined due to mining but recovered fully afterwards. Increased permeability led to enhanced well yields, but water quality deteriorated, probably because of oxidation and mobilization of in situ sulphides during the unconfined and recovery phases. At the other site, heads in a poorly transmissive sandstone fell rapidly during subsidence and did not recover; hydrogeological responses varied at the site scale due to variations in bedrock-drift continuity. Predictions and monitoring schemes can be guided by the general conceptual model, but must consider local hydrogeological variations. Effects in shallow aquifers not in direct contact with the mine can be simulated using readily available flow models.

Abstract This paper describes how tracer tests can be used in flooded underground mines to evaluate the hydrodynamic conditions or reliability of dams. Mine water tracer tests are conducted in order to evaluate the flow paths of seepage water, connections from the surface to the mine, and to support remediation plans for abandoned and flooded underground mines. There are only a few descriptions of successful tracer tests in the literature, and experience with mine water tracing is limited. Potential tracers are restricted due to the complicated chemical composition or low pH mine waters. A new injection and sampling method (‘LydiA’-technique) overcomes some of the problems in mine water tracing. A successful tracer test from the Harz Mountains in Germany with Lycopodium clavatum , microspheres and sodium chloride is described, and the results of 29 mine water tracer tests indicate mean flow velocities of between 0.3 and 1.7 m min −1 .

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. The data show that, following cessation of pumping, mine water recovery follows an exponential curve similar to the recovery of an aquifer following a pumping test. Several previously unpublished 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. 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. In coal mining areas where no water-level recovery data are available, methods for the prediction of mine water inflow and recovery modelling are proposed and the problems associated with mine water recovery modelling are discussed. The paper concludes that modelling of mine water recovery, based on mine water inflow and estimated void space, can be used to give reasonably accurate predictions of recovery times and flows, but that water level monitoring is essential for precise predictions. The control of mine water during the period when coal mining was a nationalized industry was generally based on a safety first principle. 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. The result of this policy was a general lack of experience of mine water recovery and continuing doubt about underground connections. 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. 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 void-space was calculated using roadway dimensions for supported excavations and a figure of 10% of the original extractions thickness for unsupported (total extraction) workings ( National Coal Board 1972 ). Using this principle it was assumed that mine water recovery would proceed as a series of steps, with very little recovery when water was ‘filling’ a large void, followed by a period of more brisk recovery until the next large void was reached. 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, at least at large scales, mine water recovery follows precise exponential curves that appear to be independent of the distribution of mining voids. These curves are very similar to the recovery curves observed following an aquifer pumping test.

Abstract In the UK, the first longwall faces at Wistow Mine in the Selby Coalfield, with only 80 m depth of cover to the base of the Permian, experienced several inrushes of groundwater derived from the overlying Lower Magnesian Limestone, causing serious disruption to coal production. Subsequent decrease in panel width coupled with increased depth of cover to the Permian reduced the incidence of water problems. However, 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, a major aquifer of regional importance. This opportunistic study fills that gap. Precautionary observation boreholes had been drilled above and around the margins of two proposed longwall panels prior to working the 2.5-m thick Barnsley seam at a depth of 550–600 m. Data loggers permitted continuous monitoring of the Sherwood Sandstone and Drift piezometric levels over a 2-year period. Widespread drawdown and recovery effects due to intermittent groundwater abstraction from a nearby factory were observed. Standard aquifer pumping test analyses of the hydrographs allowed transmissivity and storativity to be determined before, during and after mining. The results showed apparently permanent post-mining transmissivity increases of up to 149% around the margins of the panels, and up to 234% directly over the first panel. Post-mining storativity remained mostly unchanged. However, the greatest effects were noted during the closest approach by the second longwall panel, which also caused some additional subsidence over the first panel, when peak transmissivity increases of 1979% and storativity increases of 625% occurred. Anomalous, intra-cycle, 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. The results are consistent with similar investigations carried out at relatively shallow depths (<220 m) in USA coalfields. However, the Selby study shows that mining at much greater depths still has a significant impact on shallow aquifers, with implications for enhanced aquifer recharge, abstraction well yield and possible increased contaminant transport rates.

Abstract The recent closure of the South Crofty tin mine, the last working mine of this type in Europe, has raised questions over possible environmental consequences. During several centuries of operation, the mine was dewatered by a series of pumps located at different levels in the mine. Older workings near the ground surface were dewatered by a series of adits that discharge into nearby rivers and streams. The quality of water draining from these shallow workings is generally good and no treatment is required. However, because of bad experiences at the nearby Wheal Jane tin mine, 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. In order to address this problem and make predictions of the timing and volume of the discharge, computer simulations using the SHETRAN/VSS-NET model have been carried out. This model has already been applied to the simulation of groundwater rebound in several UK coalfields. However, the hydrogeological characteristics of coal mines differ considerably from the South Crofty mine. In this mine, the country rocks comprise granite and metamorphic slates, strata that have negligible transmissivity and very low storativity. Most of the groundwater flow is therefore in the ‘drives’ and stopes from which tin ore was extracted. The inflows to the mine during its operation were mapped and quantified, and were found to be mainly head-dependent. These inflows usually originated from fault zones in the rock, and also from nearby disused and flooded workings, which surround the modern mine. 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. The model was then used to predict the dates when surface discharges could be expected to commence. One of the main limitations of the predictions was a ‘blank’ depth interval in the mine plan records, which could be interpreted in two possible ways: (i) the zone was worked in the mid-nineteenth century but the plans were lost; or (ii) the zone was never worked. Local professional opinion favoured interpretation (i), 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. In the event, (ii) above appears to have been the true case, resulting in a marked steepening of the rebound curve during the later stages of rebound, with surface discharge commencing in November 2000, as much as a year in advance of most other predictions. 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.

Abstract The strata within and above the South Nottinghamshire Coalfield dip gently towards the east. There are many abandoned shallow workings in the western area where the coalfield is exposed, but to the east the coalfield is concealed beneath Permo-Triassic strata. The coalfield has yet to suffer closure, mine water rebound and the acid mine drainage (AMD) cycle. A very large area has been exploited with complicated internal drainage systems dependent on the maintenance of existing pumping regimes. An evaluation of the AMD threat has been carried out with particular regard to the risk posed to the Sherwood Sandstone aquifer, which overlies the concealed part of the South Nottinghamshire Coalfield. The evaluation has been assisted by three-dimensional (3-D) visualization that has enabled lumping of plentiful mine abandonment data, and predictive runs using the University of Newcastle GRAM model. These studies indicate that the critical spill-over elevation is 41 m above Ordnance datum (aOD), and that the aquifer will be at risk about 20 years after pumping ceases from the Coal Measures.

Abstract Concurrent dewatering and slope depressurization operations have been underway at the Escondida open pit since 1996. A hydrogeological investigation has been undertaken as part of the depressurization operation on the north wall. It has shown that there are two distinct hydrogeological units in the slope, the altered and unaltered Escondida Porphyry. The hydrothermally altered Escondida Porphyry is clay rich and has relatively high matrix permeability. The underlying silicified, or unaltered porphyry, has very low matrix permeability and groundwater flow is in steeply dipping NW-trending fracture systems. 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. To investigate effective drainage measures for the silicified porphyry, a groundwater model of the slope was constructed. 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.

Abstract Hydrogeological information is crucial to the development of a sound environmental 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.

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.

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). 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. Further, it may present a risk to the environment through the development of poor quality water with elevated concentrations of metals, metalloids, sulphate and depressed pH. The prediction of future pit lake water quality within a MPL is, therefore, essential in considering environmental impact on a closed or abandoned mine facility. The controls on a MPL will vary over time, and will involve chemical, biological and physical processes. Localized and regional-scale processes affect these in turn. Consequently, in order to predict pit lake water quality it is essential to understand the hydrogeological, geochemical and limnological processes that influence water quality.

Abstract Within the framework of a study on the impact of a mine-tailing spill at Aznalćollar, SW Spain, we investigated the oxidation of pyrite and other sulphides by means of two column experiments and reactive transport modelling. The columns were filled with pyritic sludge mixed up with a sandy and a clayey soil, respectively. The columns were located outdoors for 15 months and leached 10 times. Prior to simulating reactive transport, a flow model permitted a detailed description of the behaviour of the column at a daily time-scale. The most important parameter extracted was the hydraulic saturation. This parameter controlled the amount of O 2 that could diffuse into the soil, which, in its turn, affected the rate of pyrite oxidation. The sandy and clayey columns behaved very differently. In the sandy column, pH dropped due to the oxidation of pyrite. As a result, silicate minerals dissolved, providing Na and/or K that precipitate together with Fe and SO 4 as jarosite. The high concentration of Zn in the leachates was consistent with the concentrations predicted from sphalerite oxidation. The low As and Pb concentrations, however, were explained by their coprecipitation in the jarosite. In the clayey column, the dissolution of dolomite kept the pH high, impeding the dissolution of silicate minerals and precipitating amorphous Fe(OH) 3 in the place of jarosite. The model also permitted rate laws proposed in the literature for pyrite oxidation to be discussed. We found that the oxidation of pyrite by Fe 3+ was not faster than by O 2 , contrary reports in the literature. Finally, the model was used to predict the behaviour of other soil types and other sludge contents. According to the predictions the dissolution of jarosite was very important to maintaining the pH at a value of approximately 2, even for gravels or low reactive sand.

Abstract At present there is no suitable method to predict either the longevity of contaminant sources within spoil heaps, or the evolution of their strength over the contaminating life time of the sites. Existing acid-base accounting techniques provide little information relevant to the prediction of field contaminant concentrations and time scales. For robust prediction, the relative rates of contaminant generation and attenuation must be evaluated then extrapolated to the physical scale and environmental conditions of real field sites. Laboratory unsaturated column experiments on colliery spoil from a well-documented site in County Durham have been set up to assess its contamination potential, and the results compared with results from a mathematical model for contaminant release and transport. The random walk method, a form of particle tracking, is used to transport iron and sulphate ‘particles’, released by oxidative weathering of pyrite minerals. The model also includes the oxidation of ferrous iron to ferric iron in an attempt to account for contaminant ‘sinks’, for example where ferric iron spontaneously precipitates as ferric oxyhydroxide and is effectively removed from the transport process. In general, the modelled results compare favourably with the laboratory results and any discrepancies can be accounted for.

Abstract The San José Mine is a mothballed Ag-Sn mine near Oruro on the Bolivian Altiplano. A groundwater risk assessment has been carried out considering: (i) the current mine water pumping operation; (ii) potential future mine flooding; and (iii) mine waste leachate, at risk sources. Mine flooding rates have been simulated using two models (MIFIM and MODFLOW), with input data based on the observed water inflow distribution and calculated mine volumes. Mine water chemistry has been characterized by field analyses. 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. Empirical and modelled data suggest that no risk is (or will be) posed to Oruro’s public water supply wellfields at Challapamapa. Continued pumping and discharge of mine water poses a potential risk to surface water recipients and private groundwater abstractions located alongside these.

Abstract Direct observations made during underground hydrogeochemical surveys of abandoned lead-zinc mines has highlighted the precipitation of secondary zinc minerals within abandoned lead-zinc mine workings in the north Pennines. Chemical analysis of mine waters has shown that molar concentrations of sulphate exceed those of zinc by two or three orders of magnitude, although they are released in equimolar proportions following the weathering of sphalerite. The excess of sulphate over zinc indicates that there must be significant sinks for zinc within the mine workings. Secondary zinc mineral sinks (principally hydrozincite and smithsonite) are the most likely explanation for the deficit in molar zinc concentrations and these minerals have been identified underground. In addition to the secondary zinc minerals, secondary calcite and aragonite from the workings have also been shown to provide sinks for zinc (by coprecipitation and solid-solution incorporation of zinc in these minerals). Calculation of the molar quantities of zinc and sulphate involved showed that as little as 5% of the zinc, weathered daily from the mineral deposits within the workings, is found to leave the mine dissolved in the mine water. However, this is sufficient to adversely impact the ecology of the receiving waters of the River Nent, which currently receives five circumneutral zinc-rich mine water discharges and drainage from a disused aqueduct.

Abstract The Cleveland Ironstone Field (NE England) is a major sedimentary iron orefield in which the principal ore minerals are iron silicates (berthierine, chamosite) and carbonates (siderite). The siderite in this area is known to be rich in Mg and Mn in solid solution with Fe. Although this ore assemblage would not normally be expected to give rise to acid mine drainage phenomena, a number of discrete ferruginous mine water discharges (totalling some 6.5 million litres (Ml) day −1 ) have been identified flowing from abandoned underground mine workings and old spoil heaps in the ore field. Some of these discharges are extremely acidic (pH ≥ 3.3), with total Fe reaching 1220 mg l −1 . At the point of first emergence to the surface, most of the discharges are so highly charged with dissolved CO 2 that they effervesce. Upon degassing, one of the discharges precipitates a ferroan calcite deposit, which is most unusual as a mine water discharge precipitate. All the other discharges precipitate more usual ferrihydrite and goethite ‘ochres’. The geochemistry of these waters supports the view that oxidation of pyrite in the roof strata initiates dissolution of siderite in the old workings, releasing CO 2 , Fe and Mg to solution. This chain of reactions results in these waters having SO 2− 4 as their major anion (from pyrite weathering) with Mg as the major cation (except where Fe exceeds Mg in concentration). Two of the near-coastal discharges have Na as the major cation, and elevated Cl concentrations, suggesting a sea-water component. However, SO 4 /Cl ratios suggest that sea water can account for no more than 20% of these waters. Most of the Cleveland mine waters have significant environmental impacts, ranging from ecological damage to receiving water courses to flooding problems caused by the clogging of surface sewers with mine water precipitates. A range of remedial measures are proposed.