Most of the waste management problems faced by uranium mines are not unique; they are problems common to most of the mining industry. The presence of radioactivity, however, is a complicating factor. The three uranium isotopes, U-238, U-235 and U-234, found in natural uranium are parts of two radioactive decay chains, headed by U-238 and U-235. These two chains contain 14 and 11 principal radioisotopes respectively, with four and three minor radioisotopes, due to branches in the decay chains. Ultimately these chains decay into stable lead isotopes.
Although the presence of radioactivity places additional constraints on the management of the wastes, the presence of other elements such as arsenic and nickel can place greater limitations on waste management.
Waste Rock
In most cases ore is found buried in rock of no commercial value. The waste rock is the material that must be mined in order to get at the ore. In the case of an open pit mine, the amount of waste rock which must be removed to reach the ore may be far larger than the actual volume of ore: so-called stripping ratios (waste to ore) may be 40:1 or higher. However, surface mining is relatively cheap, because very large trucks, power shovels and loaders can be used, and most of the capital and operating costs associated with underground mining are avoided.
Underground mining is generally more selective, only opening narrow tunnels to reach the ore. As a result, waste to ore ratios may be less than one. The costs of underground mining are much greater, because shafts and hoisting equipment or ramps must be developed to gain access to the ore; ground support is an important safety issue; mine ventilation is vital, particularly in uranium mining; and the confined space generally permits only smaller mining equipment. The decision on underground versus open pit mining is based on safety and environmental factors (such as the stability of the rock, the depth of the ore below the surface, the presence of surface water), and on economics (weighing high unit cost, low volume underground mining against low unit cost, high volume surface mining).
Waste rock may be barren of the mineral of interest, or it may contain trace quantities at too low a concentration to be economically extracted. Waste rock may also contain other minerals which have no commercial value but which may have properties that require some degree of control for environmental reasons. For example, arsenic is frequently found accompanying some precious metal ores, and copper is found with gold. In the case of uranium, waste rock may contain trace quantities of uranium mineralisation at too low a grade to be considered ore. Other minerals may also be present. At Olympic Dam in Australia, the ore contains economic concentrations of both copper and uranium. In the case of Key Lake in Canada, both nickel and arsenic are present, but not at concentrations that have justified their recovery.
One other factor may also require consideration in both uranium and other metal mining: the presence of sulphur minerals, most notably pyrite and chalcopyrite. These minerals oxidise under weathering, with the oxidation sometimes promoted by bacterial action, resulting in the formation of sulphur oxides, which mix with water to form acids, such as sulphuric acid. Acidic waters can be very detrimental to local biological systems. The acids will also dissolve some of the heavy metals from the waste rock, increasing the toxicity of seepage from the waste rock.
To put some dimensions on the waste rock problem, consider that an underground mine may generate less than one tonne of waste rock for every tonne of ore produced. However, the quantity may still be substantial if the ore grade is low and tonnages are high. On the other hand, an open pit mine may generate 40 t of waste rock for every tonne of ore. At Eldorado's Beaverlodge operation in Canada (Ref 1), which was in production from 1953 to 1982, the underground mines produced 9 773 000 t of ore, averaging 0.21% U, and 4 300 000 t of waste rock. In contrast, the open pit mines at Key Lake from 1982 to 1997 produced 3 671 593 t of ore averaging 1.95% U and 79 620 000 t of waste rock.
As the grade of the ore increases, the quantity of rock to be mined decreases. Thus, McArthur River, a high grade underground mine currently under development in Canada, is expected to produce from current reserves and resources 1 256 400 t of ore averaging 12.75% U, and about 1 115 000 t of waste rock. At Key Lake, 2 730 000 t of the waste rock had sufficient mineralisation to require some special handling, while at McArthur River 215 000 t are expected to fall into this category.
Historically, mining operations have piled waste rock on surface, used it for back-fill of mine openings, and even used it for construction material, such as road building on the mine site. As awareness has grown of the potential problems created by acid generation and metal leaching, efforts have been devoted to studying the properties of waste rock before the mine development. In the past virtually all the drilling activity was directed at determining the grade, quality and extent of the ore zone.
Modern mining devotes as much attention to the waste rock (Ref 2), drilling the waste rock to determine the various rock types and the quantities of each. Tests are commonly done to assess the acid generation potential of waste rock and to assess the solubility of any trace mineralisation that could give rise to environmental problems. These tests generally start with a complete elemental analysis to determine the scope of possible problems. Next the acid generation potential due to the presence of sulphide minerals and neutralising potential due to the presence of carbonates are measured, giving the net neutralising potential. Accelerated weathering tests are done in humidity cells, where the rock is sprayed with water but kept unsaturated to reproduce environmental conditions in a waste rock pile above the water table. Column tests are done to reproduce flooded conditions, where waste rock is stored underwater.
Based on the results of these tests and the information from the drilling programme, waste rock management plans are developed. Rock with particularly undesirable environmental properties is used for purposes, such as mine back-fill, that will remove it from the active surface environment. In cases where the volume of problematic rock is more than can be accommodated in this manner or the rock properties are not suitable for the purpose, rock may be placed underwater, for example in a flooded, mined-out pit, to prevent oxidation and the subsequent release of contaminants. At Key Lake, about 60% of the mineralised waste will be processed with the high grade McArthur River ore, permitting the economic recovery of some uranium which would otherwise be lost. The remainder, and the McArthur River mineralised waste, will be disposed of in a flooded mined-out pit, thus removing it from the surface environment and potential problems due to weathering.
Less problematic waste rock may be left on the surface but graded to promote the run-off of precipitation and covered with a surface layer of less permeable soil, such as glacial till, and re-vegetated. The objective in this case is to reduce infiltration of precipitation and, thereby, reduce the amount of leachate that could be produced by the waste rock. Barren waste rock, which would not generate any undesirable leachate, may be left exposed to the elements with any decommissioning requirements dictated only by aesthetics. Depending upon the climate and the locale, re-vegetation of the rock pile may be indicated.
Tailings
The separation of metals from ores may be loosely broken down into two types of processes, pyrometallurgy, or smelting, and hydrometallurgy, or wet chemical processing. Uranium has generally been extracted by the latter type of process. The first step is to reduce the particle size of the ore, increasing the surface to volume ratio, to expose more surface area to the processing chemicals. This is accomplished by crushing and grinding. The desired particle size of the ground ore is dependent upon the solubility of the uranium mineral and the properties of the matrix material.
After grinding the ore to the optimum size, it is leached by exposing it to chemicals that will preferentially dissolve the uranium. In some past operations, there was sufficient carbonate in the ore to consume large quantities of acid. Because of the cost of acid, these ores were processed by alkaline carbonate leaching. However, all current uranium operations are using sulphuric acid leaching of the uranium. The uranium solution is separated from the ground rock by a series of washing stages and then further purified and concentrated by solvent extraction or ion exchange.
The impurities that have been separated from the uranium are left as a sulphuric acid solution. This is neutralised with lime and blended with the ground rock to form the tailings for disposal. Thus, tailings consist of ground rock with 5% to 20% chemical precipitate (largely gypsum and hydroxides), depending upon the exact nature and concentration of the impurities that must be separated from the uranium. They contain all the radionuclides from the two uranium chains except the uranium itself. The tailings are produced as a slurry, a mixture of fine solids in water.
Historically, the disposal method for tailings has been as much the result of local geography and climatic conditions as the requirements of environmental legislation. Digging holes costs money and many mines are operating on such a small margin that advantage is usually taken of any favourable terrain. Tailings have been placed in shallow natural depressions in the ground. Valleys have been dammed to form tailings ponds. Natural lakes have been used. Where there are no favourable surface features, tailings have been placed in constructed surface impoundments where they are allowed to drain and consolidate.
In arid regions, such as the American southwest and South Australia, the water may evaporate fast enough that no discharge and, hence, no treatment is required. In wet areas, such as northern Canada and the Northern Territory in Australia, evaporative losses are small and water management becomes of major importance in tailings disposal. In the non-uranium mining industry, the tailings themselves are often used to construct the dam which will contain them. However, for uranium tailings this may be more problematic, because the dam itself would be radioactive and may require further isolation from the environment.
Problems have been experienced in arid areas in the past with tailings impoundments that were not covered before being abandoned. Winds blew the dry tailings from the impoundment, spreading low level radioactive contamination over considerable distances. This was a significant problem at some sites in the American southwest (Ref 3). Unlined surface tailings impoundments have also resulted in seepage that has contaminated groundwater systems. However, a properly designed surface impoundment with a liner and a drainage system, and with provision to cover the tailings after drainage and consolidation, can get around these problems. An example of such a system is the Key Lake surface tailings management facility (Ref 4), although it has suffered some climate-related problems. Ice layers built up during the winter can prevent consolidation and create long term seepage, if they are not thawed before decommissioning.
As with the waste rock, the tailings may contain reactive sulphur compounds and may produce acid leachate with additional heavy metal contamination, which would be undesirable from the environmental point of view. Although there were measurable concentrations of Ra-226 in the watershed, a far greater environmental problem with tailings management in the Elliot Lake region of northern Ontario was low pH due to acid generation from the tailings (Ref 5).
In some mines as early as the 1950s, tailings were used for mine back-fill. However, this has never been a complete answer. The action of crushing and grinding the ore increases the volume, as does the addition of the chemical precipitate, with the result that a greater volume of tailings is produced than the opening left by removal of the ore. In addition, some structural integrity is often required of mine back-fill for safety reasons. This has required separating the tailings into a coarse fraction (the sands) and a fine fraction (the slimes). The sands are suitable for back-fill, but the slimes require some other management technique. Mixing cement with the tailings to add strength is another option, but this also increases the volume of material to be managed.
Earlier uranium deposits were frequently of low grade, and consequently the concentrations of contaminants in the tailings were low enough that surface deposition generally did not present serious environmental hazards. Some of the deposits now being mined are of much higher uranium grade, demanding greater care. However, the high grade means that a much smaller volume of material must be disposed of and much greater containment can be achieved without inordinate cost.
Again, to give some dimensions to the problem, consider the three mines mentioned previously. At Beaverlodge, the production of 21 236 tU resulted in the creation of 10 100 000 t of tailings, 40% of which were used as back-fill with the rest being placed on the surface, primarily in a natural lake in a low-flow watershed. By the completion of milling of Key Lake ore, total production will be 71 611 tU, resulting in about 4 400 000 t of tailings, about two-thirds of which have been deposited in the original surface tailings impoundment, with the rest being placed in the Deilmann pit. The current reserves and resources at McArthur River are expected to produce 160 200 tU and about 4 400 000 t of tailings, to be placed in the Deilmann pit.
The last 15 years have seen an evolution in tailings management thinking. Tailings manage-ment efforts in the 1970s were directed at complete isolation of the tailings, preferably above the water table, using impervious barriers, which are difficult to achieve, lock water in, and could lead to long term high pore-water pressures. Pore-water pressure is a driving force for expelling contaminants from tailings. To avoid this expulsion of contaminants, one objective in tailings management must be the elimination of excess pore-water pressure by the time of decommissioning.
For surface disposal, weathering is clearly a factor that can result in the dispersion of tailings. If weathering can be eliminated, then water becomes the only carrier of contaminants. Another objective then is to reduce water flow through the tailings; if there is no water flow, there can be no water transport of contaminants. Elimination of water movement leaves molecular diffusion as the only remaining driving force to disperse contaminants. Molecular diffusion depends only upon the concentration gradient and is a very slow process. To reduce flow through the tailings, the principal design requirements are low tailings permeability and a low hydraulic gradient, i.e. a small groundwater driving force, across the tailings. The means of achieving this is dependent upon the local conditions.
Clearly, it is desirable to get tailings out of the surface environment, which avoids erosion, and avoids human intrusion. Returning the tailings to the mine means returning them to the same environment as hosted the original ore. Disposal under water greatly reduces oxidation of the tailings, reducing the potential for acid generation, and gives further insurance against weathering and intrusion. These considerations lead to placement of tailings below the water table in mined-out pits.
The question of loss of contaminants from the tailings still has to be addressed. Flowing water tends to take the path of least resistance. The task then becomes to place the tailings in such a manner that moving groundwater tends to flow around, rather than through, the tailings. However, as tailings are placed in a pit, the mass of tailings at the top loads the system so as to increase the pore-water pressure above the normal hydrostatic head. This excess pore-water pressure must be relieved before decommissioning, to prevent the expulsion of contaminated pore water. Two similar systems are being implemented at three sites in northern Saskatchewan to accomplish the objectives of creating a relatively impermeable mass of tailings with no excess pore-water pressure within a highly permeable groundwater system (described below). The flowing groundwater will take the path of least resistance and flow around the tailings rather than through them.
Rabbit Lake
Key Lake
Two other considerations influenced the design of the Deilmann system. First, the Deilmann pit was much larger than the Rabbit Lake pit. As the tailings level in the pit rises, successive layers of tailings will become very thin and prone to freezing during winter operation. The ice lenses will impede consolidation and cause continuing seepage as they eventually melt. Second, the sandstone walls of the Deilmann pit are much more prone to weathering than is the basement rock of the Rabbit Lake pit. Over an extended period of time this could result in hazardous conditions for workers in the pit. A solution to these potential problems was to flood the pit so that the water (and ice) cover would prevent freezing of the tailings and also protect the pit walls from weathering.
The optimum system for the Deilmann pit in terms of release of contaminants proved to be a two-stage system, with initial sub-aerial deposition (i.e. without flooding the pit), similar to the Rabbit Lake system. This system is now in operation. In the second phase, the top of the pervious surround will be sealed with a layer of tailings and the pumps will be throttled back to allow partial flooding of the pit. Sub-aqueous deposition of thickened slurry tailings will be done by tremie-pipe injection into the material already in the pit (Figure 2). A deep-well thickener has been installed at Key Lake to produce higher density tailings for the sub-aqueous injection. The tremie pipe is a standard civil engineering technique for underwater concrete placement and its use has been demonstrated in tests conducted in one of the Key Lake effluent ponds and in the Rabbit Lake pit.
McClean Lake
Decommissioning
Unlike the Rabbit Lake and Deilmann ore bodies, the Jeb ore body was not under a lake. Hence, when tailings deposition is completed, the tailings will be covered, the pit will be completely back-filled, and the surface will be re-vegetated. However, the restored water table will cover the tailings, providing an additional radon barrier.
It is expected that some short term flushing of contaminants from the waste rock piles around the Deilmann pit will occur, leading to higher nickel concentrations in the pit water late in the operating period and early in the decommissioning period. Contaminated water will be pumped from the pit and treated in the mill or in the dedicated water treatment plant. The pumping will maintain the groundwater gradient toward the pit, continuing the collection of all contaminants. Water treatment is expected to be necessary for only a few years after shutdown. Biological methods of improving water quality are also being investigated as a possible passive approach for decommissioning with less human involvement.
Modelling
Samples of Key Lake tailings were characterised by scanning electron microscopy, x-ray diffraction, grain size determination, mineralogy, and tailings and pore-water chemistry. Geotechnical testing was performed to determine specific gravity, sedimentation characteristics, consolidation properties, and permeabilities. ACCUMV (Ref 10), a finite difference computer program for analysing one-dimensional, self-weight consolidation of accreting layers of compressible materials, was used to model the consolidation of the tailings. The model was validated by using it to predict the settlement of tailings in the Rabbit Lake pit. After it was shown that the program was conservative, i.e. actual consolidation exceeds the predicted consolidation, it was used to predict the consolidation of Key Lake tailings in the Deilmann pit under both sub-aerial and sub-aqueous deposition. As a further check on the modelling, three column tests of the Key Lake tailings were done, two using a 200 mm diameter by 1.5 m high column and one using a 1.0 m diameter by 11 m high column. These tests also demonstrated that the computer modelling was conservative (Ref 11).
Geochemical testing (Ref 12) was done to characterise the tailings pore water, the materials to be used in construction of the facility, the effects of residual mineralisation of the pit wall rock, and the effects of any materials which could be co-disposed with the tailings (various types of mineralised waste rock). Prior work had been done on the geochemical characteristics of the waste rock piles around the Deilmann pit (Ref 13).
Both local and regional hydrogeological modelling (Ref 14) was done using MODFLOW (Ref 15), a three-dimensional, finite difference code developed by the US Geological Survey. The accumulated knowledge of 15 years of hydrogeological monitoring and operation of dewatering systems at Key Lake supplied the input parameters for the model. Contaminant transport from the tailings is primarily by diffusion, which is dependent upon the concentration gradient between the tailings and the surrounding materials.
Other contaminant sources considered were the small residual groundwater flow through the tailings, the expulsion of pore water during the relief of the excess pore-water pressure above the hydrostatic head, leaching of residual mineralisation in the pit walls, and changes in groundwater chemistry due to the oxidation of the various geological units around the pit during the time when the cone of depression of the groundwater was maintained by the dewatering system. In addition to the tailings cases, leachate from the waste rock piles around the pit was also considered.
The analyses were carried out for three cases, full side drain (pervious surround), partial side drain, and no side drain. Although the full side drain case resulted in the least flow through the tailings, it did not result in the lowest impact. The full side drain provides a pathway by which groundwater sweeps the diffusive loading from the tailings out of the envelope and into the pond above the tailings, resulting in a higher mass loading on the environment. The no side drain case resulted in greater flow through the tailings with a corresponding higher mass loading. The partial side drain case proved to be the optimum for overall environmental impact.
The data from the regional groundwater model were used in the Environmental Transfer Pathways (ETP) model to predict impacts to local and regional human communities and ecological receptors (Ref 16). Because there are no communities in the Key Lake area, two hypothetical communities were developed: a seasonal hunting and fishing camp on the shore of Lower Key Lake, and a permanent community living on country food on the shore of Russell Lake, a larger lake downstream of the project. Calculations were carried out for natural uranium, Ra-226, Th-230, nickel and arsenic over a 10 000 year period, using both the realistic best estimate and the hypothetical worst case for environmental loadings.
In all cases, the predicted radiation doses to the receptors were far below any criteria which could be applied (worst case dose <30 µSv/year, realistic case <5 µSv/year). The water quality in the Deilmann pond was predicted to meet the Saskatchewan Surface Water Quality Objectives in all cases except the worst case Ra-226, which was predicted to marginally exceed the objective (0.12 as against 0.11 Bq/litre).
However, there were several very conservative assumptions made in these calculations. No credit was taken for source depletion as contaminants are leached from the tailings over the long term. For the purposes of the analysis, it was assumed that the pit was filled with tailings from the processing of Key Lake ore. In fact, McArthur River ore, which will be the mill feed for the last 15 to 20 years of operation, is very clean, containing only tiny traces of nickel and arsenic. Arsenic and Ra-226 are the limiting contaminants from the Key Lake ore. The result is that the pore water in the blended tailings will have a lower concentration of arsenic than the pore water in straight Key Lake tailings, and the environmental impact will be proportionately lower. In addition, the Key Lake area is not very productive and it is doubtful that there is enough game to permit a community to exist on country food, meaning that the radiation doses would be even lower than predicted.
Field Evaluation
The Rabbit Lake impact assessment was based on pore-water chemistry for fresh tailings. From sampling of pore water in the pit it has been found that, as the pore water ages, more radium precipitates out of solution, reducing the concentration in the pore water and reducing the flux of radium from the decommissioned pit. In modelling the Deilmann project, the realistic case used this fact, while the worst case assumed that the pore water concentrations would remain at the original high levels. The expectation is that the Deilmann system will work as well as the Rabbit Lake system, meaning that there is another layer of conservatism in the performance predictions.
Industrial Waste
The industrial waste from a mine-mill operation may be handled in a variety of ways, depending upon the particular material. Laboratory wastes and waste reagents are generally treated along with the tailings and disposed of by that route. Combustible wastes are generally incinerated. Non-combustible material, primarily metallic scrap, must be decontaminated if it is to be released from control. If the site is not too remote, it may be worthwhile decontaminating the material (generally accomplished by washing) and recycling the clean scrap metal to, for example, a steel mill. If the site is very remote or the decontamination process is too difficult, the material may be disposed of in the mine during decommissioning or buried in the waste rock pile.
Waste Water
Radium, specifically Ra-226, is generally the most critical radionuclide in waste waters associated with mining and milling of uranium. In some cases, where the ore contains significant quantities of natural thorium (Th-232), thorium-chain radionuclides may be important. Although the standards for radium are very strict, it is readily detected because of its radioactivity at concentrations far lower than would be possible by conventional chemistry. It is also removed from water to these very low levels by co-precipitation with barium, a stable element with similar chemistry.
Where arsenic and nickel are present in the ore, these are often more of a problem than the radioactive components, and the chemical processing of the water to remove these can be more difficult than removal of the radioactivity. Any sludges produced during the water treatment are disposed of with the tailings. To the extent possible, water is cleaned up sufficiently to allow it to be recycled in the process, rather than introducing fresh water. In arid climates, where there is little surface and groundwater flow into mines and where evaporation rates are high, it may be possible to close the water balance, so that there is no need to discharge waste water at all. However, in wetter climates the mine and the tailings management system generally produce more water than can be recycled, in which case the waste water must be treated to meet discharge limits for all contaminants before being released to the environment. The water discharged to the environment is monitored to ensure that it meets all discharge requirements before being released.
Radon
Radon-222 is the radioactive noble gas which arises as a decay product of Ra-226 in the U-238 decay chain. Because radon is not reactive and because it has a half-life of 3.8 days, it can escape from tailings and waste rock and become airborne. Rn-222 decays into a short chain of radon progeny (Po-218, Pb-214, Bi-214 and Po-214), which are particulate and generate radioactive aerosols. This is the source of the radiation dose that has been implicated in the lung cancer recognised as an occupational disease among early uranium and some other metal miners (Ref 18). Concern has been raised periodically over the population dose arising from radon generated by uranium wastes. The United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) has produced estimates of the population dose arising from this source (Ref 19). This topic is the subject of another paper in this Symposium, but some observations can be offered here.
Radon is ubiquitous in the environment. Virtually all soils and rocks contain trace amounts of uranium, which in turn gives rise to radon release to the atmosphere. Thus, low concentrations of radon can be measured in air samples anywhere in the world, although the concentrations are lower over the centres of the oceans far from any landmass. Routine monitoring around uranium mining and milling sites has shown that the radon concentrations at mining property boundaries are indistinguishable from natural background levels over land. In the province of Saskatchewan, which is the source of all current Canadian uranium production, measurements have shown that the radon concentrations are actually lower at the boundaries of the northern mine sites than they are in the southern agricultural regions of the province. The reasons for these differences have not been well investigated, but they could be related to the continued land disturbance by agricultural activities releasing soil gas with high concentrations of radon into the atmosphere.
Clearly, the contribution of uranium mining to total radon release to the atmosphere is small and any risk associated with this is hypothetical and unlikely to be demonstrable. In contrast, some of the other environmental impacts of improperly managed uranium tailings can be clearly measured. SENES Consultants (Ref 20) has shown that UNSCEAR has substantially overestimated the health risk due to radon from uranium operations. Cohen (Ref 21 and Ref 22) has questioned the validity of the linear no-threshold hypothesis in assessing these small incremental risks, which fall within the range of natural background. Recently the International Commission on Radiological Protection (Ref 23) has recognised that the use of collective dose, unrestricted in space and time, for small incremental doses arising from waste disposal activities can lead to the misapplication of resources. In any event, the currently favoured approaches to tailings disposal will result in tailings being placed at considerable depth in the ground or under water. Both methods will result in virtually no radon release after the system is decommissioned, thus eliminating this source of public dose.
Summary
The wastes from uranium mining are similar to those from other metal mining, with waste rock and tailings being the principal wastes. Although uranium wastes contain radioactivity, stable contaminants are frequently of greater concern to the environment. Uranium mines, like other metal mines, are paying much greater attention to waste rock management and the application of good science now, to avoid the acid drainage problems of the past.
Recent uranium developments have favoured the disposal of tailings in mined-out open pits, with special construction to eliminate excess pore-water pressure and provide groundwater pathways around the tailings. This approach removes the tailings from the active surface environment, reduces the chance of future human intrusion, controls the release of contaminants to acceptably low levels, and virtually eliminates any release of radon from the tailings to the atmosphere. However, a convenient mined-out pit may not always be available and a properly designed surface facility can serve very well.
At Rabbit Lake a new facility was started in 1984, placing the tailings in the mined-out Rabbit Lake pit, using the pervious surround system (Figure 1) (Ref 6). A drift, or tunnel, with a 2.2% negative slope was mined from the bottom of the pit to beyond the pit rim, where it connected with a bored raise leading to the surface. The drift was back-filled with crushed rock and a system was installed to pump water from the pit and return it to the mill for use in the process. A crushed rock drain was installed in the pit bottom and continued up the sides of the pit, with an additional liner of sand. Neutralised tailings are placed in the pit and the pore water is drained off through the pumping system, promoting consolidation of the tailings. Complete containment is achieved during operation, because the groundwater is drawn into the pit and collected through the pervious surround and pumping system.
At Key Lake the mined-out Deilmann pit was selected for the new tailings repository (Ref 7). The Rabbit Lake pit was largely in the Precambrian basement rock; the Deilmann pit is partly in the basement and partly in the Athabasca sandstone. The Rabbit Lake basement rock is fairly tight but with a high hydraulic gradient due to the adjacent high mill hill, and hence relatively high groundwater flows were expected in the basement rock. At Key Lake, there is a much lower hydraulic gradient in all geological units, higher permeability in the sandstone, and very high permeability in the sand overburden, with the result that more than 99% of the flow is in the upper sands. Prudence suggested the use of the pervious surround at Rabbit Lake to ensure a low hydraulic gradient by creating a free-flowing path around the tailings. Deilmann was ideally suited to the tailings plug concept, whereby a high density, low permeability tailings deposit is developed, which, combined with the natural benefits of a low hydraulic gradient and high by-pass flow, results in very low flow through the tailings, without the need for a pervious envelope.
The pervious surround system was initially proposed for the McClean Lake project (Ref 8), using the Jeb pit, which is the first of the McClean Lake ore bodies to be mined. However, when the processing of higher grade ores from other properties was considered, some modifications were indicated. The use of a pervious surround system requires regular work in the pit to raise the surround, as the tailings level rises. Higher ore grades would result in radiation fields in the pit that would present problems for the workers in the pit. Hence, the flooded pit concept with no surround is preferred (Ref 9). No work in the pit is required and the water gives good shielding against the potential high gamma radiation fields. The Jeb pit also has favourable hydrogeology, to permit the use of this system without the pervious surround.
The closure procedures for all three systems are similar in intent, if not in execution. After all the tailings have been deposited in the pit, a cover of sand or till will be placed on top of the tailings and the pumps will be shut off allowing original water tables to be re-established. The cover serves two purposes: initially it adds weight to consolidate the final layers of tailings, squeezing out pore water, and in the longer term it provides a diffusion barrier between the tailings and the pond. The cover material and its optimum thickness will be determined late in the operation of the tailings system, because it will depend upon local conditions at the time (for example, past experience has shown that a clay fraction in the cover material would tie up radium, if this were a problem in the pond water at the time of decommissioning).
There has been extensive theoretical modelling of all three in-pit tailings disposal systems. However, only the work on the Deilmann system will be described here, as an example of the type of attention paid to tailings disposal.
Because the first phase of tailings placement in the Deilmann pit only commenced in 1996, it is a bit early to draw firm conclusions about the actual performance. However, the Rabbit Lake pit has been used for tailings for almost 14 years. Instrumentation to measure settlement and pore-water pressure was installed in the Rabbit Lake pit in 1987, allowing comparisons to be made between predicted and measured performance (Ref 17). The tailings are settling somewhat faster (Figure 3) and the excess pore-water pressure is dissipating somewhat faster (Figure 4) than predicted in the modelling used in the original Rabbit Lake assessment. This means that the system should be ready for decommissioning with a shorter delay after the end of tailings deposition.