*Bruce W. Downing, M.Sc., P.Geo. is Senior Geologist, Gamah International Ltd., Vancouver.

 

INTRODUCTION

A significant amount of investment hinges upon the results and interpretation of Acid Rock Drainage (ARD) data. In the process of saving money spent on ARD work, QA/QC is generally limited or not done at all. A measure of confidence of the data must be ensured, and this can be obtained only from a good quality assurance/quality control (QA/QC) program. ARD studies are essentially geochemistry programs whose reliability and integrity depend upon the quality of field sampling and laboratory analytical procedures. A well conceived orientation survey will provide the background for a large scale program. A good QA/QC program will add costs to the ARD study, but it will also enhance the confidence of regulatory agencies and reviewers in assessing the data. Geoscientists and engineers may have differing interpretations of, and conclusions from, a given set of data, but an appropriate QA/QC program should minimize inconsistencies and uncertainties within the data itself. The project that may have resulted in a correct interpretation with incorrect data, will generally end up paying for it later on in the project life. Reporting engineers and geoscientists must also have some assurance that the data reported, and on which conclusions are based, is valid to the extent that it may stand up to an audit. The end product should always be questioned with "how defensible is the data ?"

Each sample should be analyzed initially using a standard Acid Base Accounting (ABA) method together with an ICP 30 element (aqua regia digestion) analysis for trace elements and a major oxide (whole rock) analysis. An inorganic carbon analysis should be considered as a indication of carbonate abundance, as carbonates are generally the predominant neutralizing material and represent readily available and reactive neutralizing capability. ICP 30 element analysis of the leachate from Neutralizing Potential (NP) tests may frequently assist in the interpretation of mineralogical data.

This analytical package will generally allow for interpretation of chemical and mineralogical characteristics of potential ARD generation and neutralization, and permit the identification of samples for which kinetic testwork is required. For each of these analytical procedures, QA/QC is extremely important. Good analysis (and good data) can be used by many disciplines.

The objectives of a Quality Assurance/Quality Control (QA/QC) program are:

METHODOLOGY

An outline of one example of QA/AC methodology is shown in Figure 1. Topics that should be addressed in QA/QC analysis are sampling, Acid Base Accounting (ABA), petrographic study, kinetic testwork, databases and reporting, chain of custody and QA/QC data analysis.

Figure 1: General Schematic of ARD Testwork Determinative Procedures with QA/QC for a Porphyry Copper Deposit

Sampling

The major and most important component of any ARD program is sampling, which may be straightforward for existing waste rock dumps and tailings, or for diamond drill core, but challenging in underground development situations (Photograph 1).

Photograph 1: Sampling a Lithology - Underground at Windy Craggy, BC.

Inappropriate or inconsistent sampling procedures produce unreliable data and can lead to erroneous interpretations. Field material sampled should be divided into categories such as overburden, waste rock (includes pit walls and pit floors) and tailings. All samples taken for ARD studies must be prevented from oxidation after sampling by enclosure in sealed containers, ideally under nitrogen. Wet samples must be dried immediately (< 40 °C), or (ideally) collected under nitrogen and frozen. The scopes of characterization and analyses for each category are very important in the initial stage of data acquisition. All lithologies should be sampled, both barren and sub-ore in order to:

  1. Determine limits of ABA for acid generating and non acid generating material.
  2. Determine spatial differences in lithologies
  3. Generate enough data points for waste rock block modeling (Downing and Giroux, 1993). Waste rock is generally defined by the Project Mining Engineer.

There are usually few limitations for the size and location of samples taken from existing waste rock piles or tailings accumulations. However, where samples must be selected or taken from drill core, as is the case for a new mine or the extension of an existing mine, there are usually rigid constraints on both the quantity and location of samples available for ARD studies. During the evaluation phase of a potential new mine there are heavy demands on available drill core for analytical and evaluation studies, metallurgical testwork, geotechnical studies and other purposes. In addition, most drilling programs are targeted at the discovery of economic ore, rather than the determination of waste rock characteristics. The selection of drill core samples for ARD studies must therefore be based on a careful examination of projected ultimate pit or underground development plans and sections. Figure 2 shows a typical section for the proposed open pit of a porphyry copper mine.

Figure 2: Pit Section and Diamond Drill Hole Location for a Copper Porphyry Deposit

Samples should be identified as to type (grab, chip, composite etc.) and sampling method(s) described. The chain of custody of samples is very important in order to identify possible errors and inconsistencies (This was clearly exemplified recently in the Bre-X case). Sampling error(s) usually will lead to cumulative errors in subsequent analyses.

Acid Base Accounting (ABA)

ABA analyses are static tests that include Neutralization Potential (NP), sulphate and sulphur analyses. There are a number of procedures available for both NP and AP (Acid Potential determined from sulphate and sulphide analyses) determination. Ideally the procedure chosen should be the most appropriate one for a given site, but regulatory agencies may specify (or highly recommend) only one acceptable procedure. It is usual practice to include the determination of Paste pH under the heading of ABA testwork. Again, acceptable procedures for this determination are usually specified by regulatory agencies. ABA analyses should include:

The potential errors associated with the highly subjective "Fizz Test" in the Sobek Method of NP determination have been discussed elsewhere in these pages.

Standards

A rigorous QA/QC testing was carried out for the ARD certified reference material NBM-1, issued in 1994 by CANMET, Ottawa (Leaver & Bowman, 1994). The reference material was supplied by Noranda from a non-ore grade pit rock from the Bell Mine in Granisle, British Columbia. An intra-laboratory cross check with 14 participating laboratories was conducted using different ABA methods. The results indicated a range of paste pH from 8.4 to 8.73, weight % sulphur ranging from 0.27 to 0.34 and NP from 48.1 to 60.2. Internal laboratory replication (i.e. standard deviation S.D.) was very good except for one lab which reported a S.D. of 4.3. Dramatic NP variations of the NBM-1 range from 15.8 to 95.2, depending upon the NP method have been reported by Lawrence. The results of the NBM-1 indicate a range of NP values and the inherent variability of one sample analyzed by several different methods. This variability of a reference sample indicates that NP values can be quite variable if different methods and/or operator deviation from the standard documented methods are used during the ARD study.

Internal site specific standard(s) should be developed for online monitoring in mining situations. This would be much more cost effective than purchasing the NBM-1 standard. The standard should be material representing the waste rock or waste rock lithologies. This reference sample should be developed at an early stage in the mine feasibility study from core or pre-stripping waste rock. The procedures of establishing a reference standard can be obtained from CANMET.

Variability of results can not be effectively judged if a standard has not be used. The question then becomes, "is it the method?" or "is it the sample?".

Kinetic Testwork

At the present time there has been little standardization of kinetic testwork procedure with regard to sample preparation, cell or column design and operation or data reporting. ASTM Designation D 5744-96, however, specifies sample preparation procedure, cell design parameters, cell operation and data reporting procedures for humidity cell testwork. In the absence of other standard procedures, the use of ASTM Designation D 5744-96 is recommended. The use of duplicate kinetic tests is highly recommended, as is the use of a "blank" or "control" using non-acid generating samples.

Petrographic Studies

The most important factor to consider is the behaviour of a mineral and its metal component(s) during weathering and their integration into the sampled medium (Borsch, 1995). Thus, petrographic studies should be conducted as part of the ABA procedure with the following general objectives:

  1. To determine primary and secondary mineralogy and alteration variation that would impact the determination of neutralization potential.
  2. To examine sulphide mineralogy.
  3. To determine modal mineralogy for rock classification as a check on field classification used in drill logs (see Figure 3).
  4. To examine grain size and grain boundaries as reaction sites for acid rock generation.

Figure 3: Graphical Representation of Modal Mineralogical Analysis

Petrographic analysis has been discussed on another page of this website. Where X-ray diffraction (XRD) is used for identification of minerals in bulk samples, individual mineral grains should be isolated under the petrographic microscope for confirmatory single grain XRD analysis.

A lithogeochemical study should be initiated as part of the ABA procedure and petrographic study with the following general objectives:

  1. To determine the bulk chemistry of the non-ore material which is a direct indication of mineralogy, including alteration.
  2. To correlate chemistry with petrographic studies.
  3. To determine weathering indexes.
  4. To determine complimentary prediction method(s) in conjunction with the conventional and regulatory NNP and NPR (NP/AP) methods (Downing and Madeisky, 1997).

Laboratories

The common laboratory errors can be classed as follows (Calow, 1991):

In this list, all except one, chemical and physical interference, have a "human component".

In conjunction with world wide standardization of laboratories, the analytical laboratories used should be registered under the ISO 4000 certification program with respect to ARD testwork. This ensures that the laboratory has good management and analytical procedures with adherence to standard guidelines and practice.

Databases and Reporting

The set-up and maintenance of a database(s) is an integral part of the study and reporting. Computerisation from the initial exploration program to the advanced stage and subsequent mining involves a large and growing database that becomes in reality a valuable asset.

What constitutes a good database and how reliable is it ? Data integrity is a constant concern. The construction of a sound and valid database begins with good sample collection. Appropriate sample collection, preparation and analytical procedures and standards must be maintained throughout a project life. Errors can be generated throughout the whole scenario of a project from data collection, preparation, analysis, input, transfer and merging through to reporting. How to eliminate or minimize errors is not the question during data analysis but how to recognise them, correct them and report them is of major importance. Check sampling and validation of the database should be carried out even though it is time consuming to the point of being 'boring'. Error recognition can be achieved through periodic printouts and plots, and/or a complete database dump followed by manual viewing and editing. This also provides a quick data reference. One should generate ways of cross checking the data through use of plots or mathematical manipulation, querying all results and basic statistics. There is always an element of luck to spot errors before final reporting. Errors always seem to crop up at the most inappropriate time. An effective way of error reduction is having the project people directly involved with the data analysis and reporting as they can best identify incorrect results. The valid database is still prone to problems when it is subdivided into sections for analysis using similar or different software, manipulations and calculations performed and the data dumped back into the original database. Retaining current database versions is very important as well as documenting the whole database. A central database must be maintained as well as doing routine backups. The end product should always be questioned 'how defensible is my data ?'

Reporting is an integral part of the study. As the reviewer of the ARD study may not be familiar with all analytical procedures, these procedures must be documented and submitted as part of the written report. Equally important are sampling procedures, sample compositing procedures and petrographic reports. Reporting of the QA/QC procedures is important so that the reader understands what was done and the results which were obtained. It is rather unfortunate that QA/QC reporting is somewhat boring and not readily appreciated by many, however it is a necessary part of data analysis.

Chain of Custody

An acceptable QA/QC program in ARD studies will require that the complete history of every sample taken for evaluation be recorded. Such a history will include:

  1. The date, time and sampling protocol for the original sample.
  2. The method, duration and location of any sample storage.
  3. A detailed record of any physical or chemical treatment of the sample, including drying, crushing, grinding, screening, splitting, and washing.
  4. A record of all personnel who have handled the sample, including time and place.
  5. Records of all disposals of sample parts, fractions and splits.

This "cradle to grave" record for a sample constitutes the chain of custody. Any engineer or geoscientist evaluating ARD testwork results may need to follow the chain of custody backwards in order to investigate unusual or unexpected results. In addition to its necessity for scientific investigation, chain of custody has important legal ramifications.

QA/QC Data Analysis

Data collected during the program should be plotted and/or statistical analysis carried out in order to determine spurious results and confidence of data. Simple plots may indicate spurious results or produce some confidence in data and interpretation, some examples are as follows:

Inorganic carbonate NP vs. NP is a method of determining whether there is any correlation between carbonate content (carbonate NP) and total NP, as determined by a specific testwork procedure. Figure 4 shows such data for four waste rock types in a large copper porphyry deposit. The data shows some correlation between carbonate NP and total NP for the Quartz Porphyry, Granodiorite and Andesite, but no correlation for the Saprolite. Under most conditions NP will be greater than or equal to carbonate NP The line of equality is shown in red in Figure 4. Exceptions to this may possibly occur where a sample contains predominantly siderite or rhodochrosite as the carbonate minerals, but this would be unusual for a copper porphyry deposit. It is a reasonable conclusion, therefore, that the single andesite and granodiorite samples, and possibly three of the six saprolite samples placing above the equality line represent faulty analyses. For those samples placing below the equality line, the degree of displacement is directly related to the proportional contribution of non-carbonate minerals to overall, or total, NP.

Figure 4: Inorganic Carbonate NP versus Total NP for Four Lithological Units of a Porphyry Copper Deposit

One extremely useful graphical method for the assessment of ABA data is a plot of Neutralization Potential Ratio, NPR (NP/AP) against sulphide sulphur analysis. By superimposing upon such a plot the critical NPR for a given dataset and for the relevant jurisdiction, and the critical sulphide sulphur content for potential acid generation, the data may be divided into four categories, or quadrants. Figure 5 shows ABA data (NPR and % sulphide sulphur) for the production monitoring samples of waste rock (two lithologies, pit foot wall and hanging wall) and tailings at the Colomac Mine of Royal Oak Mines Inc. in the Northwest Territories (NWT) of Canada . In the NWT, samples with NPR < 2 are considered potentially acid generating unless demonstrated otherwise (Steffen, Robertson and Kirsten, 1992). Samples containing less than 0.3% sulphide sulphur are generally considered to be incapable of sustaining acid generation (Price and Errington, 1995, Price, 1997). Thus the lines NPR = 2 and Sulphide Sulphur = 0.3 % are superimposed on the figure, dividing it into four quadrants. Almost all of the diorite and andesite waste rock samples fall in the upper left-hand quadrant, NPR > 2, Sulphide Sulphur < 0.3%. These samples are considered non-acid generating. A few of the diorite and andesite samples fall into the upper right-hand quadrant, NPR > 2, Sulphide Sulphur > 0.3%. These samples are also considered non-acid generating. Four of the andesite and two of the diorite samples are in the lower right-hand quadrant, NPR < 2, Sulphide Sulphur > 0.3% and would be considered potentially acid generating in isolation of the remainder of the samples. The presence of these six samples, while of note, does not affect the conclusion that the waste rock dumps will likely be, on aggregate, non-acid generating. Almost all of the tailings samples fall in a statistically tight cluster in the NPR > 2, Sulphide Sulphur > 0.3% quadrant, with an outlier group of three samples in the NPR < 2, Sulphide Sulphur > 0.3% quadrant. While these three samples may represent real values, it is possible that they may be faulty analyses. Again, the conclusion is that the tailings will be, on aggregate, non-acid generating. Depending upon the seriousness of the consequences of potentially acid generating outliers, repeat analysis may be prudent.

Figure 5: Neutralization Potential Ratio, NPR (NP/AP) versus Sulphide Sulphur for Production Monitoring Samples of Waste Rock and Tailings from the Colomac Mine, Northwest Territories (Mills, 1998 from data in Royal Oak Mines Inc., 1998)

Figure 6 shows maximum acidity potential (AP), determined from sulphide sulphur analyses, plotted against total sulphur analyses for a copper porphyry deposit. There is excellent correlation between AP and total sulphur for the granodiorite, andesite and quartz porphyry samples and no correlation between AP and total sulphur for the saprolite samples. Since all but one saprolite sample place below the trend line of the other samples, it is a reasonable conclusion that the saprolite samples contain sulphate minerals that contribute significantly to their total sulphur content, but not to their AP. However the correlation between total sulphur and AP for the other samples indicates that, for this sample suite, it would not be unreasonable to calculate AP from total sulphur, rather than sulphide sulphur using the statistical trend line from the graph. This would eliminate the necessity of sulphate analyses and reduce somewhat the cost of the ARD testwork program.

Figure 6: Maximum Acid Potential (AP) versus Total Sulphur Content for Four Lithologies of a Porphyry Copper Deposit

The above examples of data presentation and analysis are three examples of the manner in which ABA data can be examined, analysed and cross-checked for validity. In most cases the techniques used for ABA data examination and analysis will tend to be site specific in addition to following broad regulatory guidelines in most jurisdictions.

The Thompson-Howarth technique is a rigorous statistical method of calculating the differences between duplicate data in order to determine precision.

CASE HISTORIES

Case 1

A recent project conducted by one of the authors (Downing) involved two sets of ABA analyses taken approximately six months apart (1A,1B,1C,2A,2B,3A and 1A-7,1B-7,1C-7,2A-7,2B-7,3A-7). The same method and equipment were used. NP Results from the second set were very different from the first by several orders of magnitude, though the initial pH values were similar, and the rock units had not changed. The NP vs. Time data was plotted and curves were very different, Figure 7. A check with the laboratory revealed that a new person had been assigned the task for the analyses and standardized the pH titrator at 7.0 instead of pH 3.5. This accounted for less acid being consumed, hence the lower NP values. The lesson to be learned was that the person familiar with the geology, mineralogy, sampling and knowledge of ABA suspected erroneous results and was able to identify a problem which was corrected immediately by the laboratory. A duplicate sample was done, but the same error was propagated. The major fault of the author was the failure to insert an ABA standard or reference sample (or duplicate of the previous sampling in this case) into the second set, even though there were only six samples. If this was done, the laboratory would have immediately recognized a problem.

Figure 7: Amount of Acid Added (Equivalent to NP) versus Time for Two Sample Sets

CONCLUSIONS

Every ARD study must include a QA/QC section so as the reader will understand the chain of custody (or sequence of processing) and all procedures, both sampling and analytical, undergone by the samples.

A review of the QA/QC data will generally show the geochemical variability inherent in a sample as well as analytical reproducibility. There should be no unacceptable error in the data to justify re-analyzing the samples. From review of internal laboratory standards and duplicates, all analytical batches of sample data should be within acceptable limits. Geological data has an inherent "scatter" due to variability and complexity of the mineralogy.

We cannot stress strongly enough that a geologist must be involved right from the beginning as that person is most familiar with the geology and mineralogy, a necessity for the sampling and help with the interpretation(s).

 

ŠThe contents of this web page are protected by copyright law. Please contact the authors for permission to re-use the contained information.

 


REFERENCES

American Society for Testing and Materials (1996), ASTM Designation: D 5744 - 96 - Standard Test Method for Accelerated Weathering of Solid Materials Using a Modified Humidity Cell, ASTM, West Conshohocken, PA, 13p.

Borsch, L. (1995), Some Observations on Mineral Properties and Analytical Reproducibility in Geochemical Samples, Mining Engineering, June, 1995, pp.567-569

Calow, R. (1991), Quality Control/Quality Assurance in Geochemical Laboratories, Explore, no. 72, pp. 23-24.

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.

Lawrence, R., 1996, Summary Notes MEND Prediction Workshops, workshop presented by MEND technology transfer committee and MEND prediction committee, December 7-8, 1995, Noranda Technology Centre, Pointe-Claire, Quebce.

Leaver, M.E. & Bowman, W.S., 1994, Interlaboratory Measurement Program for the Standard ARD Material NBM-1 Environment Canada, Canmet, Ottawa, MSL Report 94-28 (CR).

Mills, C. (1998), A Critical Review of Geochemical Stability, Acid Rock Drainage (ARD) and Metals Leaching Potential, and Tailings Management Aspects of the Royal Oak Mines Inc. Colomac Mine Interim Abandonment and Restoration Plan (Draft), report to the Northwest Territories Water Board, Yellowknife, on behalf of the Dogrib Rae Band, Rae-Edzo, Northwest Territories, 20p.

New Canamin Resources Ltd. (1995), Huckleberry Project - Mine Development Certificate Application, April 1995 - Volume VIII: Appendix III-5: East Zone Waste Characterization Program.

Price, W.A. and Errington, J.C. (1995), ARD Guidelines for Mine Sites in British Columbia, BC Ministry of Energy, Mines and Petroleum Resources, Victoria, 29p.

Price, W.A. (1997), DRAFT Guidelines and Recommended Methods for the Prediction of Metal Leaching and Acid Rock Drainage at Minesites in British Columbia, British Columbia Ministry of Employment and Investment, Energy and Minerals Division, Smithers, BC, (April), 143p.

Ramsey,M.H., Potts,P.J., Webb,P.C., Watkins,P., Watson,J.S. & Coles,B.J. (1995), An Objective Assessment of Analytical Method Precision: Comparison of ICP-AES and XRF for the Analysis of Silicate Rocks, Chemical Geology, vol. 124, pp. 1-19.

Royal Oak Mines Inc. (1998), Colomac Mine Interim Abandonment and Restoration Plan (Draft), Submission to the Northwest Territories Water Board, April:

Steffen, Robertson and Kirsten (B.C.) Inc. and B.C. Research and Development (1992), Guidelines for Acid Rock Drainage Prediction in The North, Indian and Northern Affairs Canada, Ottawa, Ontario.

Thompson, & Howarth (1976),

Vallee, M. & Sinclair, A. (1998), Quality Control of Resource/Reserve Estimation. – Where do we go from here?, CIM Bulletin, vol. 91, no. 1022, pp. 55-57.


Return to ARD at Enviromine Homepage