You are now on InfoMine, the Internet's premier mining information site



 

INTRODUCTION

Many deposits provide challenges in resource estimation that not only include the quantity of minable minerals in a deposit but also the amount and location of materials that might have implications for environmental impact during development, operation, closure and post closure/ remediation The potential for acid rock drainage (ARD) from waste material is recognized as a key issue in project planning in the following areas:

A waste material estimate must be calculated in conjunction with the resource estimate because of the importance for mine waste disposal. ARD waste material includes overburden, waste rock, pit walls, pit floor and tailings, which should be classified as acid generating (AG), potentially acid generating (PAG), potentially acid consuming (PAC) or potentially neutral (PN) for regulatory and permitting purposes. The initial ARD study should begin at the exploration stage when management deems the deposit to have the potential to be an economically feasible operation.

The type of predictive ARD block model that is used should be decided by the resource practitioner. It will be used in long range planning and production decisions which ultimately lead to the estimation of the disposal costs associated with a block of waste rock. This is very important since each waste block must be treated separately as to its mode of removal or salvage, placement and subsequent treatment.

It is not the purvue of this paper to present a dissertation on geostatistical computations involved in block modelling, but to use the method to show that the amount of waste material should be determined in conjunction with the resource calculation and to present some ideas on methodology. Waste rock modelling has been included in at least two studies reported in the literature (Downing & Giroux, 1993, and Bennett et al., 1997). Experience from other resource studies by the authors has shown that waste rock block modelling is generally not part of any resource estimation whether at the pre-feasibilty or feasibility stages.

There are two main components of the block model:

  1. Estimate the acid generation potential of waste material, and
  2. Estimate the metal or trace element component of the waste material that would impact the metal leaching (ML) component of the ML/ARD program.

WASTE MATERIAL

Waste material is deemed to be any material that does not have a positive economic value, other than a disposal and treatment/remediation value. Potential waste material includes topsoil, overburden, waste rock and tailings.

Overburden

Overburden is a term used by geologists and engineers to designate material of any nature, consolidated or unconsolidated, that overlies a deposit of useful materials, ores, or coal, especially those deposits that are mined from the surface by open cuts. By others, overburden designates only loose soil, sand, gravel etc., that lies above the bedrock. (Thrush et al., 1968). In the northern climates, permafrost may be considered a special type of overburden, while in the southern climates, saprolite may be considered overburden. Overburden for the reclamation stage would designate the organic horizon(s) as a special overburden category as it is collectively removed as a soil salvage operation and stockpiled and treated separately for reclamation purposes.

The ARD assessment of overburden is necessary because it may be used for borrow material for roads, land fill and dams. The overburden impacts groundwater in terms of chemistry and flow. It also has a direct impact upon surface water quality in terms of water runoff and erosion contributing to sedimentation in streams. There will be an impact upon ground water quality if mineralized materials are incorporated into the overburden, depending on their particle size, quantity and distribution.

Waste Rock

The ARD evaluation of waste rock is necessary in order to assess any potential acid generating waste rock that will have to be monitored and stock-piled. Some of the non-acid generating rock may be used for construction material for roads, land fill etc. It should also be noted that glacial ice may be considered as waste rock (Downing & Giroux, 1993), with the advantage that it disappears with time.

Waste rock will have been determined based on grade cutoff value(s). The lithogeochemical difference between the hangingwall and footwall waste rock lithologies must be investigated and a distinction made when estimating the impact on the waste rock ARD block model.

Pit Wall and Pit Floor

ARD assessment of the pit walls and pit floor is necessary as they will have an impact upon water quality during mining operations and post-mining, until the pit eventually fills with water. The upper exposed walls will not be covered and have the potential to generate ARD if they contain any acid generating material. Closure and post closure remediation plans generally include modelling the pit filling and pit lake chemistry (Bursey et al., 1997)

Tailings

The ARD assessment of tailings is necessary because they will have an impact upon the method of tailings disposal, whether it be surface, subaqueous or submarine impoundment. The reserve estimate in conjunction with the mineral process flowsheet will give an estimate of the amount of tailings that will be generated by mining. This is critical in designing the proper tailings storage facility. Another component of ARD assessment is the ABA testwork to determine the potential for tailings acid generation.

SAMPLING

Sampling procedures for the deposit consists of surface and underground drilling, surface and underground chip sampling, underground and surface bulk sampling for metallurgical testing and minor amounts of trenching. It is imperative that the deposit be properly sampled so that reliable estimates can be made according to industry standard definitions of reserve and resource. There is no finite number of samples for ARD studies, but the question should be will the number of samples be sufficient in order to construct a waste rock block model that will characterize each waste block with some confidence". This may necessitate carrying out waste rock drilling in order to have sufficient sampling to properly characterize the outlying waste blocks and provide some confidence in extrapolating to other less sampled areas using an appropriate interpolation method. Additional holes will generally have to be drilled in order to define waste rock boundaries and for sampling in order to produce data for block modelling.

Core logging is vital to the block model process. Lithological, mineralogical, structural and geotechnical data are entered into the site computer database. The geotechnical data has important ramifications for pit wall modelling. Specific gravity (SG) should be measured on site using core samples regardless of rock type , so as to calculate material mass. There will be some variation in specific gravity measurements due to variable lithology, mineralogy and inclusions. Particular variants include amounts of sulphides and magnetite in waste rock. It is imperative that an accurate and detailed geological and assay - geochemical database be established for resource studies and geological drill logs should be designed to incorporate greater detail in conjunction with sampling for ARD

Quality assurance/quality control and data verification should be carried out in accordance with good management practices (see Downing & Mills, 1998) .

BLOCK MODEL

To produce a meaningful block model, containing estimates of ARD values, a data base of ARD parameters, similar to a resource estimate data base, is required. Visual and mineralogical observations together with analytical data can provide an effective means of predicting which waste rock types will be AG, PAG, PAC and/or PN. From ABA studies and geological data together with laboratory and field weathering tests, it should be possible to predict acid generation characteristics of waste rock throughout the deposit. A major problem with any waste rock block modelling is the lack of availability of data points in waste rock, as most drilling is involved with delineating ore. Drilling specifically to delineate waste rock is generally not an accepted method at the feasibility stage, however it is vital data that is needed for a comprehensive waste material management plan. During most drilling programs, the waste rock is never sampled, and as such data points for block modelling are not available. Generally, if the whole drill hole is assayed, then sufficient meaningful data points may be available for modelling.

Major elements have been investigated as to their potential use as an inexpensive surrogate method in lieu of neutralization potential (Downing and Madeisky, 1999). One of the first studies was the use of Ca % vs. NP plots in the estimation of potential acid consuming (PAC) material for the waste block ARD model from the Windy Craggy massive copper sulphide deposit (Downing & Giroux, 1993). The NP was attributable to carbonate mineralization, hence, leachable Ca was analyzed using the 30 element ICP-MS method following an aqua regia digestion. Day (1995) used Ca + Mg concentrations determined by ICP (following an aqua regia digestion) as a surrogate for NP. An ARD orientation study using Ca and Mg analyzed using the ICP-MS method would provide the necessary information as to whether the NP is attributable to carbonates. If a correlation does exist, then all samples should be analyzed using the ICP method since this would also provide some QA/QC to NP determinations. The ICP analysis is much cheaper than NP determinations, thereby allowing many more samples to be characterized as AG, PAG, PAC and/or PN . Generally, an ICP instrument is not found in most mine laboratories, however Ca and Mg can also be adequately measured using an Atomic Absorption emission spectrometer which many mines have on site. If a correlation can be made between CO2 and Ca, and assuming that CO2 is indicative of carbonate and that there is no graphite present , then a Leco carbon analyzer can be used. An orientation survey needs to be conducted in order to show the regulatory people that these methods have merit and is applicable to the minesite ARD testwork program.

A block model is three-dimensional spatial representation to quantify the geology and economics of a deposit which in turn is used in mine planning, see Figure 1 and Figure 2. Blocks may be of uniform size (ie., 10 x 10 x 10 metres) or they may be of variable dimensions that depend on the mining plan. The limits for the block model should be chosen to include the proposed pit outlines plus additional areas for pit enlargement(s) and to cover possible mineralization beyond the known deposit limits. The actual waste blocks that would be removed are dependant upon the mining plan. As more closely spaced data becomes available the model can be refined.

The first step in building a block model, for any variable, is to produce a detailed geologic interpretation that can be converted into a three dimensional geologic solid model. The geologic solids should be built by the geologists most familiar with the project, as this step often entails subjective decisions on boundaries and contacts of individual geologic units. The procedure is usually accomplished by interpreting geologic contacts from drill hole core logging, on cross sections cut through the deposit. The interpreted contacts are digitized, using any one of numerous software packages available. The interpreted contacts, limits of geologic domains and drill hole traces are then plotted on level plans cut at reasonable intervals, often the proposed mining bench height, from the topographic surface down to the lower limits of mining. The contacts and domain limits are drawn on each level plan joining the interpretations from each cross section. This is often a difficult process that requires a high level of subjectivity, hence the reason for involving the geologists most familiar with the project. This stage of geologic modelling is normally completed for any resource/reserve estimation and may not require much, if any, change to use in an ARD study.

The next step, would consist of reclassifying drill hole data, in terms of the interpreted geologic model. Often small intersections, of a particular rock unit, will be incorporated within larger solids of a different rock unit, during the modelling stage. Thus down hole geologic drill logs should not be used to determine limits of grade composites. Instead, drill hole data should be compared with the three dimensional solids, and the entry and exit point of each drill hole through each solid should be recorded. The subsequent data file containing hole number, from, to and geologic unit code can be used to form uniform down hole composites that truly honour geologic boundaries. For an unbiased estimate, all data should be composited to a uniform support (ie. composite length +/- one half the composite length). At the boundaries of each geologic unit, small composites, less than one half the composite length, may have been formed. These small intersections should be recombined with the adjoining sample, to avoid the loss of any assay data.

Once composites, for the variables to be estimated, have been produced, they should be evaluated as a function of rock code. Simple statistics, consisting of mean, standard deviation and coefficient of variation and histograms, should be produced, for each variable, and compared. Geologic domains exhibiting similar distributions, can be combined.

Given sufficient data to model, variography can be used to predict anisotropy and ranges of continuity, for each variable. Semi-variograms should be produced, for each variable, within each geologic domain. Semi-variograms allow for the consideration of a sample's position in space, in addition to it's distance from the block during the weighting of samples in interpolation. A knowledge of range of continuity and anisotropy, if present, also aids in designing infill drill patterns.

The final stage of producing a block model for ARD characteristics, consists of interpolating a value into each block, for each of the variables. The geologic code for each block, from the block model, determines which composites and which model will be used for the estimate. For example, in the Windy Craggy case history, (Downing & Giroux, 1993) calcium was kriged for each block in the model using ICP values, present in most assayed samples, while sulphur was estimated by the inverse distance squared method. Sulphur, in this case, was only available in ARD samples and those taken for specific gravity (2,999 samples with sulphur analyzed compared to 21,416 sampled with calcium values). As a result, semi-variograms for sulphur could not be produced for all the various geologic domains.

The criteria for characterizing waste rock as AG, PAG, PAC and PN are site dependant and may be arrived at from static and kinetic testwork and regulatory permitting If several block models have been tried, then the criteria for classifying waste blocks might be two out of three ratios to meet the acceptable PAG, PAC or PN values. Each block may have various components composed of PAG, PAC and PN material.

Data that is needed for modelling waste rock must include total sulphur, CO2 (short term buffering effect), Ca and Mg and NP.

Once the model and subsequent ABA database have been constructed, it should be continually updated from the [daily] blast hole ABA sampling. It is imperative that onsite ABA analytical work be conducted. The optimum ABA sampling procedure and analytical method(s) would be carried out in the initial months of mine operation.

CONCLUSIONS

Waste rock modelling must be an integral part of a geostatistical resource estimate study. The ability to characterize each waste block as AG, PAG, PAC and/or PN in the block model is of importance for mine planning and waste rock disposal. This information can also be used in predictive water flow chemistry models.

A waste rock ARD prediction, disposal and inventory scheme should be established by knowing the potential resource. This scheme must take into account the mine geology, blast hole cuttings (rock types, carbonate content by acid and staining tests, estimated sulphide content) and analytical results of drill and blast hole data. The scheme operates on input from the mine geologists who are responsible for the geological field input and decision(s) together with the analytical results from the mine assay laboratory. A computerized decision model can be built based upon rules, hypotheses and common sense leading to logical relationships or concepts.

The site waste rock ABA data must be reconciled with the modelled ABA of the waste blocks in order to determine how well the original prediction was and that the modelled waste blocks are correct. This reconciliation should be done on a yearly basis.

ACKNOWLEDGEMENTS

This paper is dedicated to the late Chris Mills who spent many, many hours generating this web site, and who would have been a co-author.

BIBLIOGRAPHY and REFERENCES

Bennett, M.W, Kempton, H.,J, and Maley, J.P., 1997, Applications of Geological Block Models to Environmental Management, , Fourth International Conference on Acid Rock Drainage, Vancouver, pp.293-303.

Bursey, G.G, Mahoney, J.J., Gale, J.E., Dignard, S.E., Napier, W., Reihm, D. and Downing, B.W., 1997, Approach Used to Pit Filling and Pit Lake Chemistry on Mine Closure - Voisey's Bay, Labrador, Fourth International Conference on Acid Rock Drainage, Vancouver, pp.257-275.

Day,S., 1995, Summary Notes MEND Prediction Workshops, Dec. 7-8, 1995, Noranda Technology Centre, Pointe-Claire, Quebec.

Downing,B.W., and Giroux,G., 1993, Estimation Of A Waste Rock ARD Block Model For The Windy Craggy Massive Sulphide Deposit, Northwestern British Columbia, Exploration and Mining Geology, Vol 2., No.3, pp. 203-215.

Downing,B.W., and Madeisky,H.E., 1997, Acid Rock Drainage Study of the Voisey's Bay Ni-Cu-Co Massive Sulphide Deposit, Newfoundland, poster at International Conference Acid Rock Drainage, May, 1996.