Classification of Mineral Resource Estimates

by | Apr 13, 2016

resource classification pitfalls

In the first of a series of articles on best practice DOs and DON’Ts, Bill Lewis, senior geologist with Micon in Toronto, highlights the need to tread carefully when classifying Mineral Resources.

The classification of Mineral Resources is an important factor in the analysis of risk on advanced mining projects, and implementing best practices in this area can improve the probability of a successful mine development.

A Qualified Person (QP), who is assumed to have the necessary experience relevant to the subject matter, determines the resource classification for each mineral resource estimate. However, in reviewing and auditing a large number of resource estimates over the years it has become apparent to Micon that, notwithstanding the CIM definition standards, the continuity of geology and grade are not always reasonably or sufficiently implied, assumed or confirmed to support the classification of the mineral resource estimate. Also, grade continuity can be more important than geology when criteria related to continuity are applied to the classification of mineral resources and this is often missed during the process.

Examples of some questionable resource classification practices include the following:

  • Resource blocks or mineralization outlines that are based upon a single drill hole.
  • Small numbers or groups of blocks that are given a higher or lower classification than surrounding blocks.
  • Use of geostatistical procedures without sound knowledge of the technique. Blocks are often classified using poor variography or based upon variography without regard to further geological or supporting information. This also applies to the classification of resources based solely upon geostatistical information, including categories that are based upon confidence limits, that does not also consider either the continuity of mineralization or the geology.
  • Situations in which there are strips of higher confidence-level resources between sections (where enough data can be seen from the adjacent sections), with a strip of lower-confidence resource on the section (which does not see all the way to the adjacent sections).
  • Historic resources translated into current resources and reserves without regard to the quality of the supporting information or with no supporting information.
  • Resources translated into higher confidence categories when historical mining has occurred, despite there being limited or no information regarding the location of the historic mining stopes or adequate records of the amount of material previously mined. This, in some cases, may lead to overestimation of the resource tonnages as well.
  • Classification of resources where drilling recovery is questionable. i.e. less than 70% core recovery, or selective recovery of chips or mineralization due to water issues.
  • Absent or uncertain down-hole surveys in long holes, which should result in down-grading the resource classification.
  • Lack of geological reasoning on the part of the QP estimating the resources

The first two points were demonstrated recently when a QP made the following statement regarding his resource estimate for a vein-style deposit, in which Measured, Indicated and Inferred (MII) resources formed concentric rings around the drill hole intercepts: “I don’t see any problems with having a “spotted dog”* picture of MII. MII is not an indication of the ore body continuity … but [of] the distribution of the drill holes intersecting it, which will make whatever pattern that it does. As a matter of fact, for a spotty, discontinuous deposit, pinching and swelling, it kind of makes sense to have the MII look the way I had it.”  We disagree.

Micon has seen several similar cases in which a circle is drawn around a single drill hole to indicate Inferred resources.  This style of interpretation misses the point that resources should be based upon “the location, quantity, grade or quality, continuity and other geological characteristics of a Mineral Resource [being] known, estimated or interpreted from specific geological evidence and knowledge, including sampling”.  A solitary drill hole does not evidence geological and grade continuity, but only reflects that mineralization is present at that point.

Another of the most common issues is that isolated blocks are given a higher classification than surrounding blocks, and vice versa.  Such cases are usually the result of isolated blocks being estimated in the first or second passes of the interpolation alogrithm, based upon the drill-hole spacing.  This is a fairly common issue which can lead to a “spotted dog” effect of resource blocks within an estimate.  The way to resolve this issue is for the QP to review the data in the context of the geology and mineralization and reclassify the isolated blocks accordingly. Isolated blocks are sometimes easily identified because the surrounding classification takes on the same outline as the search ellipse.

This problem becomes critical when a company is pressed for time and/or unwilling to complete further work to increase the confidence in the mineral resource estimates.  This was the case when Micon reviewed a draft feasibility study (FS), in which the owner did not want to complete a new mineral resource estimate even though considerable diamond drilling had been completed since the last estimate.  The draft FS used the previous mineral resource estimate and stated that “85% of the Diluted Resource was converted to Ore Reserves, of which 32% was from the Inferred category”.  When confronted with this issue, the company stated that it could not produce a mining plan without relying on the inferred resources, due to the spotty nature of the Measured and Indicated resources based upon the older resource estimate. The rework this created resulted in a significant delay in completion of the FS, which was issued almost 12 months later.

Resource blocks classified using poor variography, or based upon variography without regard to further geological or other supporting information, is another common issue.  This also applies to any classification of resources based upon geostatistical information, or confidence limits, that does not also consider either the continuity of mineralization or the geology.

Micon has noted numerous examples of this during reviews of resource estimates.  For instance, Micon’s review of a mineral resource on a Latin American project resulted in a large proportion of the Inferred resources being upgraded to Indicated.  The lower confidence in the original estimate arose because the author classified the resource based upon his geostatistical study of uncertainty, without recognizing that the deposit had clearly demonstrated continuity of width and fairly consistent grades over at least two kilometres of strike length.  This was subsequently confirmed when infill drilling (50 holes) agreed with the predicted grade and width of the mineralization 96% of the time.  In this case, the geostatistical methods applied caused the resources to be classified at a lower category than it should have been.  While not all of the Inferred material could be upgraded to Indicated, the majority of it could, based upon the demonstrated continuity of grade, width and strike length of the mineralized zone.

In contrast, on another recent project, the mineralized zone exhibited extreme variations in the grade over short to very short intervals in all directions within the mineralization envelope. Although, in some cases, the mineralization was continuous over longer intervals, the mineralization generally exhibited poor grade continuity, either in its width or along the strike of the primary zone of mineralization.  Within the zone of mineralization, 67% of the contained metal was located within the top 10% of the percentile (decile) assay data distribution, with 44% of the contained metal in the top 2 percentile and 37% in the top 1 percentile.  While the global resource tonnage and grade estimated by Micon were comparable to the original estimate, the tonnage and grade classified by Micon in the Measured and Indicated categories were significantly lower because of the low continuity of grade within the mineralized geological envelope.

Another problem Micon has encountered occurs where strips of higher confidence blocks occur between cross-sections or levels, while the blocks on the sections or between the levels themselves have lower confidence.  One such situation involved a North American mine exploiting a single vein.  In this case, the Measured resources were projected approximately 10 metres from the channel sample point, with Indicated resources projected another 15 metres from the Measured resources, and Inferred projected an additional 10 metres beyond that.  This led to an interpretation in which strips of Measured Indicated and Inferred resources were formed between working levels one above the other.  An additional complication was that only the channel samples from the levels were used to inform the estimate, while the drilling information was ignored.  Micon’s review, led to a change of practice for the operator, wherein both the channel and drill core samples were used to estimate the resources and, where mining was being undertaken, the material was classified as either Measured or Indicated.  This resulted in a significant increase of the Measured and Indicated resources which could then be added to the mine plan as reserves.

Micon has encountered numerous examples in which historical estimates were converted into current resources, and even reserves, without regard to the quality of the supporting information or with no supporting information.  Perhaps the most extreme case was a mine in North America in which many of the underground workings had been dormant for several years and most were flooded. Records supporting the historical polygonal estimates, showing the actual estimation work on the cross-sections, could only be located for fewer than half the historical blocks.  For the remaining resource tabulations there were no supporting data.  Nevertheless, the owner released the historical resource estimates as a current mineral resource, without verification of the data, and published the results in a NI 43-101 Technical Report. These resources were then used as the basis for current mineral reserves without, accounting for dilution or losses.  Micon’s review of the source information also revealed that, where records existed to back up the estimates, the reserve classification was not supportable, and this led to a number of reserve blocks being down-graded. In some cases the drawings indicated that subsequent mining had reduced the remnant block size.  On sections where the historical estimates could not be verified, these resources were all down-graded to an Inferred classification.  In addition, smaller remnant blocks (less than 1,000 tonnes) were eliminated, until such time as an engineer could verify that their extraction would be economic, given that the necessary infrastructure to access these blocks would have to be re-established.

Micon has noted several other examples of resources being classified into higher confidence categories, without making proper allowance for the fact that extensive mining had occurred and there was limited or no information regarding the location of the mined out stopes or the quantity of material previously mined.  In one extreme case a project located in South America, with a history of gold mining dating back to precolonial times, had surveyed or sampled less than 5% of the historical workings.  With the exception of a few sporadic production records and surface evidence of at least 200 adits, no accurate information existed as to the extent or actual location of the underground workings.  Furthermore, artisanal miners continued to work on the property, with few records being kept of their activities.

A North American company was attempting to secure rights to the deposit and create one large mine. In addition to purchasing the ground from landowners, it conducted a drilling program to determine the extent and grade of the mineralization.  In several instances, the drill broke through into old stopes or workings. Only a provisional estimate of the tonnage and grade of the “missing” material was possible. Accordingly, Micon was unable to classify any of the resources at higher confidence than an Inferred category. Later, though, based upon the same information a multi-million ounce resource estimate including Measured and Indicated categories was made, classified using “distance… to the nearest composite used in the estimation.” There was no attempt to quantify the extent of underground workings on the veins, other than those few adits that had been channel sampled.

Lack of or suspect down-hole surveys for long holes should likely result in a lower classification of the mineral resource estimate.  This will generally occur in projects with a lot of historical drilling for which down-hole surveys are absent or which only used dip tests to determine the inclination at specific distances down the drill hole but without factoring the hole’s azimuth.  Micon has encountered a number of such projects over time and usually recommends that the client drill either twin or new holes to confirm the location and orientation of the mineralized zones, prior to classification of any resources beyond the Inferred category.  Confidence in the location of each sample is one of the most important factors in building a block model, hence it is critical to obtain the best possible survey of the drill collars and down-hole trace.

Apart from historical drill holes, Micon has noted two other survey problems that appear to be prevalent.  First, early stage exploration drill holes are often not surveyed properly until the delineation stage commences.  Consequently, part of the resource derived from the early discovery holes has to be down-graded due to lack of confidence in the mineralization geometry.  Second, in some instances, magnetic influence prone survey instruments have been wrongly used to survey drill holes in highly magnetic geological environments.

Lack of geological reasoning is evident where a resource is defined on the basis of a single drill hole. A mineral deposit and its resource estimate are three dimensional objects.  Therefore, the data needs to account for strike length, width and depth.  Trenching may determine strike length and width but not depth, whereas a single drill hole at depth will provide the width and depth but not the strike length, only the possible strike direction.  In addition, a single hole does not meet the standard that the “geological evidence is sufficient to imply but not verify geological and grade or quality continuity” necessary for an Inferred resource, let alone any higher classification.

A general rule of thumb for classified mineral resources to meet Measured, Indicated and Inferred would be for new drilling intersecting mineralized blocks to meet the following criteria:

  • Measured Resources; the addition new drilling or information should not change the overall interpretation or grade by more than 10% to 15%.
  • Indicated Resources; the addition of new drilling or information should not change the overall interpretation or grade by more than 20% to 25%.
  • Inferred Resources; the addition of new drilling or information changes the overall interpretation or grade by 25% or more.

The above guidelines rely on the premise that continuity of geology and grade has been clearly demonstrated prior to classification of the resources.  The exception to the above rule of thumb would be bonanza style gold and silver deposits where the nugget effect can have a significant effect on the grade.  However, in these cases, while the grade may vary widely, the continuity of the high grade shoots still needs to be clearly demonstrated.

All of the above issues regarding classification of mineral resources at a project can be addressed and corrected if close attention is paid to how the different classifications are applied to the estimate based upon the supporting information and a thorough review prior to issuing the estimate to the public. Also, prior to publication the question to be asked is whether or not the classified resources really address the spirit and substance of the definitions rather than trying to address preconceived notions of what they should be or how much of the resources should be in each category.

Micon welcomes your comments and feedback related to this topic. Watch this site for more articles highlighting best practice in other topics in the coming weeks, too.

*Acknowledgement: the term “Spotted Dog” comes from an article by P.R. Stephenson et al. presented at the 6th International Mining Geology Conference at Darwin, NT, August, 2006.

27 Comments

  1. Julian Poniewierski

    Good article Bill. Some interesting errors pointed out. A few I’ve seen – a couple new to me. One point though. I find it interesting that geologist will quote a Measured Resource when a grade might change by 15% with extra drilling. There are many mines that a change in head grade of 15% would literally “break” – making them uneconomic. And I don’t think promotors and executives realise that is the risk being implied when using a measured resource.

    Reply
    • Bill Lewis

      Thank you for your comment Julian. You raise a very interesting point regarding your comment about the 15% change and we will endeavor to answer this in a future article.

      Reply
  2. Terry Bottrill

    How wonderful to see a comprehensive article emphasizing the importance of continuity – especially geological continuity on resource estimation. This should be published in a journal or at least in the SEG Newsletter or a similar outlet for as wide a distribution as possible. I would very much like the opportunity to be able to use it as a reference when dealing with clients who, in the past, have not wanted to listen to arguments similar to those presented herein. Well done and thank you.

    Reply
    • Bill Lewis

      Thank you, Terry. I am going to discuss continuity specifically in my next article. Yes, by all means use the article as a reference.

      Reply
  3. Peter Broad

    Good points Bill, but one of the prime reasons PEAs fail is an over dependence on Geology. It should be required of a PEA that there is a fundamental metallurgical study to assess the basic recovery and the degree of liberation (grinding) needed to achieve an economical recovery.

    Reply
    • Bill Lewis

      Hello Peter, thank you for your comments. We will endeavor to address your comments regarding metallurgical studies in a future article.

      Reply
  4. Placido Gonzales Gomez

    Buenos puntos los tomados en cuenta en este articulo, sin embargo es muy importante aplicar ala estimacion de recursos minarales la geologia estructural, estos juegan un punto muy importante, ya que muchas veces cortan la continuidad de mineral o concentran el enrequesimiento en zonas de fallas.

    Reply
    • Bill Lewis

      Gracias, Placido. I agree it is important to apply any structural geology information into the resource estimate and I will comment on this in a future article discussing the subject of continuity in general.

      Reply
  5. pankaj sinha

    Good article

    Reply
  6. Arthur R. Leger, V.P. Exploration, Atoka Gold Corp

    Very nice summary Bill,

    In my opinion the first and possibly the most important is the drill results. I have reviewed twinned reverse circulation vs core and have found extreme variations in assay results. Often the problem is hole survey – none completed thus the actual location of the hole is unknown. A second problem is the actual assay data. Who did the work and were standards used.

    In my view, drill, drill, drill to confirm mineralization and try to interpret the correct structural configuration.

    Reply
  7. Bill Lewis

    Thank you, Arthur. Yes I agree proper drilling is the best way to confirm mineralization and therefore continuity.

    Reply
  8. Alexander Arizanov

    many Resource are reported as Block model, the mineralised zone is not outlined. By accepting angular shape of orezone following BM we add waste for ore, and we ore for waste. I agree with Arthur (drill, drill, drill) but many companies try to economize $ from all possible activities. The final result is more trouble.

    Reply
    • Bill Lewis

      Thank you Alexander for your comment. Yes you are correct, a lot of resource estimates are conducted without the mineralized zone being outlined and in these cases a secondary check after conducting the resources will more often pick up any issues, but there is no substitution for drilling and a good understanding of the deposit.

      Reply
  9. Abani R Samal

    Good article by William (Bill) Lewis.
    Certainly the resource classification does not depend on metallogenic epoch and provinces.
    The reference of 15% and 20% is really not a hard rule – rather it should be a guideline. The scale should also be considered. By scale I mean years of mining, such as for all material planned to be mined in next 3 years or so.
    Another issue: if all of the measured resources are converted into reserve, a grade fluctuation of 10 to 15% can cause poor reconciliation and big head-ache for the processing plant.

    Reply
    • Bill Lewis

      Hello Abani, Thank you for your comment. Yes the 15% and 20% are guidelines and scale should be considered. As for reconciliation usually anything under 10% when I was in operations was considered good. However, poor reconciliation is also not always a result of the processing plant but can be caused by other issues in the operation as well or could be due to a poor resource to reserve conversion due to inaccurately used modifying factors. However, this will be covered in future articles.

      Reply
      • Alex Arizanov

        yap, 10% looks the most common boundary, but all discrepancies need to be investigated properly without blaming in term to find out where the ‘real’ problem sits and get right decision. problem is many times people try to cover up and the cause can’t be determined properly.

        Reply
        • Bill Lewis

          Alex, thank you for your comment. I agree, when I was in operations we did our reconciliation to the mill on a monthly basis so that we were always on top of it. It also helped that management worked together without playing the blame game or cover up which is what happens in a lot of cases. Reconciliation is a very important tool but in a lot of cases is not addressed as much as it should be.

          Reply
  10. Richard Kehmeier

    Thank you for the article. It is written in plain English so that even non-Technical people can understand some of the issues of classification of resources. Interesting, you do not mention QA/QC and its consideration in the classification of resources. Often on properties with historical drilling QA/QC is missing or incomplete. I downgrade resources if QA/QC is missing even if other factors point to a higher classification because of a lack of confidence in the data.

    Reply
    • Alex Arizanov

      very true Richard. somehow people ignore this problem as it is important not only for classification but the whole Resources. I hate it when things are done for ‘market’, to make the project to look better; this happens too often.

      Reply
    • Bill Lewis

      Thank you for your comments Richard. Yes QA/QC is an issue which will be addressed in a subsequent article I am working on dealing with continuity in general in resource estimates. I also look at drilling recovery as well as the QA/QC to try to determine classification.

      Reply
  11. Alma Vega

    I agree with all but in narrow veins with historical data and using the polygonal method if change to new method using the block model is necessary classify like indicated and inferred resource and not depend if you have few or a lot samples because to clasify in measured depends if you have good QA/QC to clasify in measure resource and the distance between the drillholes comes to variography to design the drillholes mesh to add indicated resource to measured and inferred to indicated.

    Reply
    • Bill Lewis

      Thank you for your comment Alma. You are correct that if you are taking historical data and a polygonal model and converting it into a block model you have to be careful classifying it, and will need to have good QA/QC data to classify it as measured. I would add that the historical QA/QC would need to be verified prior to classifying it in any category other than inferred. Yes, the variography will most certainly assist in designing the drill spacing necessary to increase confidence in the data and thereby the mineral classification.

      Reply
  12. Eldrick Esper

    Very good article, remembering some usual and important mistakes under the resources classification, that we need always take care about how we go fix them. The continuity of the orebody is a good point to discuss and the spott dogs effect when you have just one hole far the mineralization. and ofr course a good review on the QA/QC routines will allow us a good estimation and confidence of the classification of resources.

    Reply
    • Bill Lewis

      Thank you for your comments Eldrick, I always appreciate the feed back. I agree with your points and I hope you enjoy the second article in this series discussing continuity in general. comments

      Reply
  13. Richard J Small

    A good summary of some of the many pitfalls awaiting the unwary mining geologist, although one would have hoped that they would be avoided by a ‘proper’ QP!

    Reply
    • Bill Lewis

      Thank you Richard for your comments. Unfortunately as history shows us, given the many failures, they are not always avoided by a “proper” QP. Only further diligence and care will hopefully help us and allow us to avoid the pitfalls.

      Reply

Submit a Comment

Your email address will not be published. Required fields are marked *