PROFESSIONAL TECHNICAL AND SCIENTIFIC PAPERS

 

Impact of Competent Persons' judgements in Mineral Resources classification

 

 

S.K.A. Owusu; K. Dagdelen

Mining Engineering Department, Colorado School of Mines, Golden, Colorado, USA

Correspondence

 

 

---

ABSTRACT

Uncertainty with regard to estimated grades and tonnages of a mineral deposit demands risk assessment in order to build investor confidence and attract the interest of other stakeholders in the success of a project. Uncertainties associated with Mineral Resource estimates can lead to unreliable production schedules and unpredictable cash flows. However, the techniques used in the mining industry to determine these uncertainties are inconsistent, because the important decisions taken in the process are solely dependent on the responsible Competent Person (CP), without limitations. This leads to disparities between different CPs' results, using data-sets from the same drill-holes. The various standard codes for public disclosure provide guidelines and recommendations for the classification of Mineral Resources and Reserves but do not provide details on, for example, the amount of geological and geostatistical information needed to qualify for each category of Resources and Reserves. The parameters used to generate the classification categories are subjectively assumed by the responsible CP. In this paper we investigate the impacts of different CPs' judgements on resource classification, using the same data-sets. The results underpin the need for the mining industry to develop a uniform Mineral Resources and Reserves classification framework that can minimize or avoid significant discrepancies that lead to potential misleading public disclosures.

Keywords: mineral resources, mineral reserves, classification, reporting codes, uncertainty

---

 

 

Introduction

Mineral Resource classification plays a key role in the economic assessment of mining projects, as investors typically make investment decisions based on the information used to generate the cash flow analysis. Due to inconsistencies in Mineral Resource reporting by various Competent Persons (CPs), it is crucial to investigate how different resource classification techniques are applied in the mining industry to categorize Mineral Resources as Measured, Indicated, and Inferred, based on the uncertainty assigned to each class. The Mineral Resources report should provide reliable information on the deposit under consideration and define the different Mineral Resource classes, based on the confidence levels assigned to the different blocks of the orebody model. The international standard reporting codes for Mineral Resources promote competence, materiality, and transparency in public disclosure (Shaw et al., 2006).

The principle of competence refers to a responsible, suitably qualified and experienced professional with at least 5 years of relevant experience, who is required to be a member of an organization recognized by the specific reporting code and capable of demonstrating competence among his or her peers. The principle of transparency requires that all available, accurate, and sufficient information is presented. Materiality requires the inclusion of all relevant and reasonable deposit information to enable investors and their advisors to make balanced judgements based on the information presented. To minimize or avoid misleading public disclosures, the standard reporting codes were established to encourage investor confidence in the exploration and mining business. Until the later part of 1980s, there were no industry standards for mineral asset reporting, and this led to doubtful and erroneous reports from individuals as well as companies.

The categorization of Resources and Reserves relies on the judgement of the CP in charge of a project, based on knowledge and experience, in conjunction with others if necessary. Each CP decides on the assumptions used and justifies the outcomes produced from each class of the Mineral Resources. The codes do not prescribe how CPs should conduct their assessments to classify Mineral Resources (Noppe 2014). The inconsistencies in the expected accuracy, precision, and confidence in the classifications can result in varying grades and tonnages of total Mineral Resources calculated by different CPs using the same drilling data. An effective estimate of Mineral Resources with a credible classification leads to reliable mine designs, efficient production schedules, robust business plans, and solid financial forecasts. The big question is; which CP produces a genuine (or not) Mineral Resources report? One common inaccuracy found in tonnage determination for a deposit is due to the application of an erroneous density or tonnage factor. Parrish (1993) stated that the most common error found when conducting Mineral Resources and Reserves audits is an error in the tonnage factor used to derive the tonnage of the orebody. The density used to convert volume to tons is crucial when determining the real in-situ tonnage and metal content of Mineral Resources, because an error of a few per cent in the bulk density can significantly alter the economic viability of a deposit. This is especially obvious in marginal projects, as the higher the tonnage factor, the lower the tonnage of a deposit, and vice versa. However, the methods used to determine the density considered in the estimation and classification of Mineral Resources are inconsistent.

The importance of bulk density is stressed in the various standard codes such as the Canadian Institute of Mining, Metallurgy and Petroleum (CIM) Mineral Exploration Best Practice Guidelines (CIM, 2018) and the Australasian Institute of Mining and Metallurgy (AusIMM) Guide to Good Practice (AusIMM, 2001). For some projects, there is insufficient data to adequately characterize the assigned density or tonnage factor of the waste and ore, while for others, there are good databases that contain well-documented density determinations. Some CPs in the mining industry assign different tonnage factors to ore and waste, while others assign an average tonnage factor to all rocks, depending on the nature of the deposit as well as their own judgement. Due to the subjectivity and dependency on the CP, the resources classified as Indicated by one CP may be classified as Inferred by another.

 

Misleading public reports and poor project outcomes

A deceptive Public Report on Mineral Resources due to erroneous estimation and classification assumptions can lead to poor production outcomes. For example, a publicly reported resource mistakenly classified as Indicated rather than Inferred can mislead investors and create future problems, including loss of investor confidence and lawsuits. There have been historical antecedents of some public announcements on Mineral Resources and Reserves where the expected risks and level of maturity of the projects were presented in incorrect contexts, thereby yielding undesired outcomes. A typical example of an estimation scandal that hit the mining industry is the Bre-X saga in 1997, where the company fraudulently claimed about 47 million ounces of gold in the Busang property in Indonesia (Groia, Bradley, and Jones, 2008).

In the 1980s, an investigation into 35 Australian gold mines showed that 68% failed to deliver the planned head grade (Burmeister, 1988). In North America, a review of about 50 projects found that only 10% achieved their commercial aims and 38% failed within a year (Harquail, 1991). An investigation into the start-up performance of a nine underground base metal mines In Australia established that only 50% achieved the designed production by the third year and 25% never achieved it at all (Ward and McCarthy, 1999). A study in the United States to compare the final feasibility study figure with the average sustained production rate from 60 steeply-dipping tabular deposits established that 35% of the mines were unable to achieve their planned production rates (Tatman, 2001). According to Noppe (2014), one partner in a coal deposit joint venture deal reported double the resource tonnage of the other JV partner, each using the same drill-hole data. This happened because the partner who produced the inflated estimate did not apply the likely mining parameters for the expected underground scenario.

Considering the intensive capital funding and risks of mining projects, BHP Billiton's Olympic Dam project in Australia is a typical example (Valle, 2011). The project cost was estimated to be US$27 billion, hence huge sums of money would have been lost if the Mineral Resources and Reserves were wrongly classified and the mine failed to meet the production target. In the mining industry, little has been done in terms of investigating the impacts associated with different CP assumptions and judgements applied to classify resources. In this research, we provide quantitative analyses of different CP assumptions, using data-sets for copper and gold projects.

 

Mineral Resource classification methods

Although other techniques may be used to classify mineral deposits, the two basic methods used in the mining industry for Mineral Resource classification tasks are the geometric and geostatistical techniques. The geometric method considers the amount, proximity, and location of data available to classify resource blocks, while the geostatistical method uses model uncertainty and configuration of the neighbouring data to classify blocks. Figure 1 outlines the two classification techniques.

A recent survey conducted on the different Mineral Resource classification techniques practiced in the gold mining industry showed that approximately 93% of CPs prefer the use of the geometric method, while 7% use the geostatistical method (Owusu and Dagdelen, 2019). The study was undertaken, using 45 NI 43101 technical reports filed on the System for Electronic Document Analysis and Retrieval (SEDAR) website, to evaluate the state of practice concerning Mineral Resource classification techniques for gold deposits. Public Reports from 2006 to 2018 were compiled and another publicly available general classification reporting guideline from a major gold mining company in the United States was included. The review covered 20 junior, 20 mid-tier, and five major gold mining companies in North America, South America, Asia, and Africa. Different gold deposit types were covered, including gold-copper porphyries, orogenic gold, breccia pipe, carlin-type, lode and placer, epithermal high- and low-sulphidation, and greenstone belt.

After critical analysis of the classification guidelines used by the various companies, it was established that 43 of the reports applied geometric methods and two combined geostatistical and geometric methods. These included techniques such as drill-hole spacing (DHS), distance to nearest drill-hole (DNDH), number of samples (NS), number of drill-holes (NDH), ellipsoidal search (ES), octant search (OS), drill-hole intercept (DHI), and kriging variance (KV). In most cases, some of the CPs combined multiple techniques like NS and NDH (NS + NDH), DHS and NDH (DHS + NDH), NS and ES (NS + Es). Table I and Figure 2 represent the various classification techniques applied for gold deposits used in the research. In the pie chart, the percentages are rounded to their whole numbers. After evaluating the different reports, it was found that there is currently a lack of uniformity in resource classification reports due to the different assumptions made by individual CPs. Also, it was found that industry players prefer the the geometric method due to its time-saving and simplistic characteristics. Hence, there is a need to develop easier and quicker techniques for the geostatistical method to enhance its use in the resource classification process.

 

 

Silva and Boisvert (2014) conducted a similar survey from SEDAR and after reviewing a total of 281 technical reports, found only 120 had sufficient information to determine the techniques used for Mineral Resource classification. The information used for the evaluations included classification technique employed, the chosen criteria, and the drilling pattern. Silva and Boisvert did not provide detailed information in terms of the mineral commodity types, deposit types and locations, company categories, or parameters applied in the classification techniques, among others. According to the investigators, the most commonly used method was the geometric, representing about 80% of the reports. It included the NS technique which constituted approximately 50%, followed by DHS (30%) and KV (about 6%), with the remainder (14%) being unspecified. The techniques used by the various CPs were dependent on either regular or irregular drill-hole spacings.

In most reports, regularly spaced drill-hole data was used for the DHS technique, while NS was applied to irregularly spaced drillhole data. Also, some industry professionals combined the different techniques to classify irregularly spaced data.

 

Research methodology

Classification is commonly performed on a block-by-block basis but the block volumes are chosen to be reasonably large and contiguous, because of the common perception that confidence in the grade should not change abruptly between adjacent blocks (Deutsch, Leuangthong, and Ortiz, 2016). To provide mining industry professionals with quantitative illustrations to demonstrate classification inconsistencies due to individual CP judgements and assumptions, we assessed data from a single bench at a copper deposit and from a gold deposit. Each case study included data validation, exploratory data analysis, geological modelling, variogram modelling, block modelling, grade estimation, model validation, and resource classification. Different assumptions made by different CPs in some of the SEDAR technical reports were applied to these two data-sets. The differences in the results indicate that the practice of applying individual CP assumptions without limitations creates discrepancies in the outcomes, because each CP has a strong influence on the parameters used to generate Mineral Resource classes.

 

Case study A: A single bench at a copper deposit

The bench is 90 m length χ 90 m width χ 15 m height. The data is considered as synthetic because much geological information is not provided, but it is useful for the purpose of the study. Thus, the research focuses on using the same data to analyse how different CP assumptions can produce varying outcomes. The data is from 36 irregularly spaced vertical drill-holes with a single sample from each hole.

The drill-hole information includes eastings, northings, elevations, and copper assay results. A block model of 7.5 m χ 7.5 m χ 15 m blocks was created for the data-set, in correspondence to the 15 m bench height. The available data from regularly spaced blast-holes 7.5 m χ 7.5 m yields similar statistical results to the exploration data, hence the choice of block size.

Statistical and geostatistical analysis

The distribution of copper in the deposit was analysed using statistical and geostatistical techniques. Table II and Figure 3 show the summary statistics and the histograms of the original composite data and the declustered data respectively. All copper grade units are given as percentages (%Cu). Variogram modelling was generated from the data to determine the spatial continuity of the data points.

 

 

Considering the inadequacy of the data, good variograms and correlograms could not be created. The geostatistical parameters generated from the correlogram model are shown in Figure 4.

Block modelling and resource estimation

MineSight software was used to generate the block model and the applied interpolation technique was ordinary kriging (OK). For the purposes of this study, an arbitrary tonnage factor of 0.354 m³/t (12.5 ft³/t) was assigned to all rocks. For validation purposes, inverse distance squared (ID²) and nearest neighbour (NN) estimation techniques were used to generate results for comparison. The three models are displayed in Figure 5. The swath plot is another tool used to validate block models, as it shows the moving window mean plots of block grades at different locations. Figure 6 shows how the results from the three estimation techniques were represented well on the swath plot. Table III summarizes the tonnages and grades generated from the three models at different cutoff grades.

Mineral Resource classification

Mineral Resource classification was performed to determine the Measured, Indicated, and Inferred classes. Search distances for the scenarios were calculated from different percentages of the variogram range, while the same number of composites and number of drill-holes were assigned to each category in each scenario. Table IV presents the details of the classification parameters for each resource class, using a sill range of 25.6 m (84 ft), number of composite (NC) samples used, and number of drill-holes (NDH) used.

The different classes generated using MineSight are shown in Figure 7. Measured, Indicated, and Inferred Resources are indicated by red, green, and yellow respectively. The different classification results for each scenario in terms of grade, tonnage, and metal content are presented in Table V.

 

 

Discussion - case study A

Table V shows the inconsistencies in Mineral Resource classification after the application of different parameters based on different CP assumptions and judgements. As seen in Figure 7, scenario 1 produced few blocks of Inferred class, while scenarios 2 and 3 did not produce any Inferred blocks. Also, scenario 2 produced Measured blocks for the entire deposit, with three blocks of Indicated class. Finally, scenario 3 generated more Indicated blocks than Measured blocks. Although the sample population of the single-bench data is small, the results from each of the three scenarios were generated by considering assumptions used by CPs in technical reports. This has shown that there can be overly or insufficiently assumed parameters, which can be applied by individual CPs to classify Mineral Resources.

 

Case study B: Gold deposit in California

The gold deposit, discovered in 1978, is located in California. The deposit consists of a large but low-grade set of veins, 1829 m (6000 ft) wide and 305 m (1000 ft) deep, in a fault zone. It was formed in a shallow epithermal and hot spring environment with gold and silver as the primary mineral commodities. Other associated minerals in the deposit include mercury, lead, iron, copper, thallium, arsenic, antimony, and zinc (Homestake Mining, Western Mining History, 2019). This case study considers the gold mineralization in the southern portion of the deposit (6600N - 10500N). This portion generally strikes north-northwest (NNW) and the dip varies depending on the orientation of the vein set.

Statistical and geostatistical analysis of data

Three estimation domains were generated after the statistical and geostatistical analyses of the data, as shown in Figure 8. The original axis range values were in feet (ft) and later changed to metres (m) for the purposes of international readership. The variogram models for each estimation domain are shown in Figure 9 and the parameters generated from each variogram are presented in Table VI. The classification parameters of the three different scenarios used for this investigation were created from the assumptions made by the CPs of three companies in the 45 technical reports compiled from SEDAR.

The classification parameters that were used by the CPs included percentage of variogram range for search radius, minimum number of composites, maximum number of composites, number of drillholes, and number of composites per hole. Table VII shows the resource classification parameters for the three geometric scenarios. Brief descriptions of the scenarios are as follows.

Scenario 1 represents the 2018 resource classification parameters used for Coeur Mining, Wharf Mine site technical report in the USA (Jimmerson et al., 2018). Based on the NI 43-101 technical report, OK was used for the estimation and the sample search distance for each class was determined based on a certain percentage of the sill range of the major continuity direction of the variogram. Three classification passes were used to categorize the deposit into Measured, Indicated, and Inferred classes.

Scenario 2 represents the 2008 resource classification parameters used for the technical report of Kinross Gold, Cerro Casale Project in Chile (Henderson et al., 2008). The report shows that the CPs applied ID² for the estimation interpolation and the search distance for each class was determined based on a percentage of the sill range of the omnidirectional correlogram. There were six classification passes for the three resource classes.

Scenario 3 represents the 2017 resource classification parameters applied in the technical report for Barrick Gold, Goldstrike Mine in the USA (Krutzelmann et al., 2017). The report shows that the CPs used ID² for the estimation interpolation and the search distance for each class was determined based on a certain percentage of the sill range of the major continuity direction of the correlogram. For the Measured class estimation pass, a box search of 12 m χ 12 m χ 6 m (40 ft χ 40 ft χ 20 ft) was used to include only composites found within each evaluated block. Thus, blocks without samples did not qualify for a Measured category. There were five classification passes for the Measured, Indicated, and Inferred classification categories.

At cutoff grades from 0.283 g/t to 1.417 g/t (0.010 to 0.050 oz/t), the different mineral resource assumptions applied on each estimation domain from the different scenarios produced different results. Again, this substantiates the discrepancies in Mineral Resource reports in public disclosures due to the application of different CP judgements in the estimation and classification processes. Tables VIII to X show each scenario's tonnages and grades generated from domains 1, 2, and 3 respectively at different cutoff grades.

Discussion case study B

The tonnages produced from the three different scenarios in the three domains are shown in Figures 10, 11, and 12 for Measured, Indicated, and Inferred classes respectively.

Measured category

Scenario 3 produced far less tonnage in the Measured class than the other two scenarios in all three domains. In domain 1, scenario 1 generated more tonnage than scenario number 2 at all the cutoff grades. Considering domain 2, scenario 2 produced more tonnage than scenario 1.

Indicated category

Scenario 3 produced far greater tonnage in the Indicated class than the other two scenarios in all the three domains. In domains 1 and 3, scenario 2 generated more tonnage than scenario 1 at all the cutoff grades. Domain 2 produced more tonnage in scenario 2 than scenario 1 at cutoff grades of 0.28 g/t to 0.85 g/t. At cutoffs of 1.13 g/t and 1.42 g/t, scenarios 1 and 2 produced almost equal tonnages.

Inferred category

Scenario 3 produced the greatest tonnage in the Inferred class in all the domains and at all the cutoff grades. Scenario 2 generated more tonnage than scenario 1 in all three domains.

The figures clearly display the differences in tonnages produced from the three CP assumptions applied to the same gold deposit. Converting the Mineral Resources into Mineral Reserves for this deposit after applying modifying factors, including technical, economic marketing, and governmental factors would also produce different results. The differences in the Measured and Indicated Resources shown in Tables VIII-X can have significant impacts on investment, development, and mining production decisions. This is a clear indication of resource classification inconsistencies, since the mining industry lacks a uniform classification framework. Diligent analysis of the classification results produced from the same data-sets corroborates the conclusions of Owusu and Dagdelen (2019) from their review of 45 technical reports on SEDAR. Considering the three scenarios applied on the same data and the corresponding outcomes, it is difficult to distinguish the right results from the wrong, since each CP followed the CIM best practice guidelines.

 

Conclusions

Mineral Resource classification is a critical factor in the success of a mining business, as it provides the confidence level that can be ascribed to a project. In spite of its importance, the techniques used to determine the uncertainties are applied inconsistently in the mining industry because the parameters utilized in the process are subjectively determined by the responsible CP. We have presented quantitative illustrations of the impacts of different CP assumptions and judgements to demonstrate that varying grades, tonnages, and metal contents can be generated from the same drill-hole data, leading to discrepancies in classification reports. In the 45 technical reports from SEDAR used for this work, the responsible CPs provided reasons for choosing their classification parameters and each report was assumed to be acceptable as per the CIM best practice guidelines for public disclosures.

This work has shown that the practice of applying individual CP assumptions without limitation can cause misleading public disclosures and affect future project outcomes. It underpins the need for a standard or uniform resource classification framework, with particular emphasis on an acceptable range of parameters to be applied per each deposit type, based on the available geological and geometallurgical information. An effective and acceptable uniform framework will help minimize the effects of individual CPs' assumptions on the Mineral Resource classification process and thus, enhance investor confidence in mineral projects.

 

References

AusIMM. 2001. Mineral Resource and Ore Reserve estimation - The AusIMM Guide to Good Practice. Edwards, A.C. (ed.). Monograph 23. Australasian Institute of Mining and Metallurgy, Melbourne.         [ Links ]

Burmeister, B.B. 1988. From resource to reality: A critical review of the achievements of new Australian gold mining projects during the period January 1983 to September 1987. MSc thesis, Macquarie University, Sydney, Australia.         [ Links ]

CIM. 2018. Mineral Exploration Best Practice Guideline. Prepared by the Canadian Institute of Mining, Metallurgy and Petroleum (CIM) Mineral Resource and Mineral Reserve Committee. Adopted by CIM Council November 23, 2018. Montreal, Canada.         [ Links ]

Deutsch, C.V., Leuangthong, O., and Ortiz, J.M. 2016. A case for geometric criteria in resources and reserve classification. Annual Report 08. Paper 301. Centre for Computational Geostatistics, Alberta, Canada.         [ Links ]

Groia, J., Bradley, J., and Jones, A. 2008. The aftermath of Bre-X: The industry's reaction to the decision and the lessons we all have learned. PDAC Conference, 4 March 2008. Toronto, Ontario, Canada.         [ Links ]

Harquail, D. 1991. Investing in junior mining companies. Proceedings of the 6th Mineral Economics Symposium. Canadian Institute of Mining, Metallurgy and Petroleum (CIM), Montreal, anada.         [ Links ]

Henderson, R.D. 2010. NI 43-101 Technical report for Cerro Casale roject, Northern Chile. Public Disclosure on SEDAR. Kinross old Corporation, Canada.         [ Links ]

Jimmerson, S.J., Nelson, K., Auld, T., Lindsay, E., Hoffer, M.R., Key, .K., and Lippoth, K.B. 2018. Technical report for the Wharf peration, Lead, South Dakota. Public Disclosure on SEDAR. Coeur Mining Inc., USA.         [ Links ]

Krutzelmann, H., Cox, J.J., Evans, L., and Collins, S. E. 2017. Technical report on the Goldstrike Mine, Eureka and Elko ounties, Nevada. Public Disclosure on SEDAR. Barrick Gold Corporation, USA.         [ Links ]

Noppe, M.A. 2014. Communicating confidence in mineral esources and mineral reserves. Journal of the Southern African nstitute of Mining and Metallurgy, vol. 114, no. 3, pp. 213-222.         [ Links ]

Owusu, S.K.A. and Dagdelen, K. 2019. Critical review of mineral esource classification techniques in the gold mining industry. Proceedings of the 39th international Symposium on Application of Computers and Operations Research in the Mining Industry, Wrocklaw, Poland. vol. 3. Routledge. pp. 201-209.         [ Links ]

Parrish, I.S. 1993. Tonnage factor - A matter of some gravity. Mining ngineering, vol. 45, no. 10, pp. 1268-1271.         [ Links ]

Shaw, W.J., Godoy, M.C., Miller, G., and Larrondo, P. 2006. An pproach to more objective classification of mineral resources. Proceedings of the 6th International Mining Geology Conference. Darwin, NT. Australasian Institute of Mining and Metallurgy, Melbourne. pp. 85-89.         [ Links ]

Silva, D. and Boisvert, J. 2014. Mineral resource classification: A comparison of new and existing techniques. Journal of the Southern African Institute of Mining and Metallurgy, vol. 114, no. 3, pp. 265-273.         [ Links ]

Tatman, C.R. 2001. Production rate selection for steeply dipping abular deposits. Mining Engineering, vol. 53, no.10, pp. 34-36.         [ Links ]

Valle, D.D. 2011. Olympic Dam expansion supplementary environmental impact statement. BHP Billiton, Olympic Dam ine, South Australia.         [ Links ]

Ward, D.J. and McCarthy, P.L. 1999. Startup performance of new ase metal projects. Adding value to the Carpentaria Mineral rovince, Mountain Isa, Queensland. Australian Journal of Mining, April 1999.         [ Links ]

Western Mining History. 2019. Information on Homestake Mining ompany, McLaughlin Mine. Napa County, California. https://westernmininghistory.com/mine_detail/10310645/ [accessed 25 June 2020] ♦         [ Links ]

 

 

Correspondence:
S.K.A. Owusu
Email: oaksolo@yahoo.com

Received: 22 Feb. 2021
Revised: 23 Feb. 2022
Accepted: 28 Feb. 2024
Published: July 2024

 

 

ORCID:
S.K.A. Owusu http://orcid.org/0000-0001-7709-578X
K. Dagdelen http://orcid.org/0009-0003-6650-8616