Loading ...
Sorry, an error occurred while loading the content.

AI-GEOSTATS: What are appropriate measures of reliability for Mineral Resource and Ore Reserve estimates?

Expand Messages
  • Bill Shaw
    Greetings You may be aware that in Australia the Joint Ore Reserve Committee (JORC) has called for submissions regarding the next revision of the JORC Code.
    Message 1 of 7 , Feb 25, 2002
    • 0 Attachment
      Greetings

      You may be aware that in Australia the Joint Ore Reserve Committee
      (JORC) has called for submissions regarding the next revision of the
      JORC Code. The current JORC Code (effective September 1999) can be
      downloaded from <http://www.jorc.org/> www.jorc.org and the call for
      submissions is also on that site.

      I am involved in the teams looking at the definitions for Mineral
      Resources and for Ore Reserves.

      Frequently questions are raised about the expected reliability of
      estimates for Resources classified as Measured, Indicated or Inferred or
      for Ore Reserves classified as Proved or Probable. There are a number of
      arguments, both for and against, regarding attempting to quantify the
      accuracy (and precision?) of such estimates.

      The Joint Ore Reserve Committee is considering whether it may be timely
      to introduce some discussion into the Code regarding the pros and cons
      of such quantification of expectations.

      Not all such arguments for and against quantifying the reliability of
      estimates are technical, for example:

      * FOR - The primary purpose of the Code is to ensure Public
      Reports best inform investors. It thus appears important that similarly
      classified estimates have similar “reliability” even if such reliability
      can only be broadly generalised, rather than quantified.
      * AGAINST - The JORC Code is not currently prescriptive and
      relies on the Competent Person making decisions about many issues
      including Transparency, Materiality and Competence (all defined in the
      Code). Quantification of “expected reliability” presumably requires
      estimates of tonnes and grades to be made using a method that would
      allow “errors” to be also estimated. This may make the Code more
      prescriptive if such errors are only provided by certain estimation
      methods.

      Either way, if the JORC Code discusses this issue it must get it right.
      For example it is not very meaningful to suggest that Measured Resources
      are “within +/- 10%” if such a statement can never be tested or
      demonstrated. Thus it has been suggested that confidence limits and
      parcel size must also be stated.

      We would appreciate your considered opinion on this specific issue: What
      are appropriate measures of reliability for Resource and Reserve
      estimates and how they are currently used by practitioners? I will
      undertake to collate submissions, to circulate these and to prepare an
      overview for JORC. Please circulate this email to any colleagues that
      you believe are interested in this issue.

      Regards,

      Bill Shaw
      <mailto:wshaw@...> wshaw@...




      [Non-text portions of this message have been removed]
    • Isobel Clark
      Bill There are several presentations being given this week at the APCOM in Phoenix. You might want to look at the proceedings and check through those. One of
      Message 2 of 7 , Feb 26, 2002
      • 0 Attachment
        Bill

        There are several presentations being given this week
        at the APCOM in Phoenix. You might want to look at the
        proceedings and check through those.

        One of the papers I have heard so far compares three
        criteria and shows how they affect a particular case
        study. The three are a weighted distance criterion, a
        kriging variance and an information measure. In this
        particular case, kriging variance came out bottom.

        Pity, because that's what I use!
        Isobel

        Isobel Clark
        http://uk.geocities.com/drisobelclark

        __________________________________________________
        Do You Yahoo!?
        Everything you'll ever need on one web page
        from News and Sport to Email and Music Charts
        http://uk.my.yahoo.com

        --
        * To post a message to the list, send it to ai-geostats@...
        * As a general service to the users, please remember to post a summary of any useful responses to your questions.
        * To unsubscribe, send an email to majordomo@... with no subject and "unsubscribe ai-geostats" followed by "end" on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list
        * Support to the list is provided at http://www.ai-geostats.org
      • Farquhar, James (HI)
        I m not sure that any one parameter can provide the measure. Resources can be classified on 1. drill data and support (quality of the drill logs, sampling,
        Message 3 of 7 , Feb 26, 2002
        • 0 Attachment
          I'm not sure that any one parameter can provide the measure. Resources can
          be classified on

          1. drill data and support (quality of the drill logs, sampling, assays,
          standards and duplicates);
          2. drill hole spacing (adequacy for the resource type, eg less than
          variogram major range);
          3. structural complexity;
          4. interpretation quality, strings and triangulations (how well do the
          geology strings fit the drilling and the regional geology);
          5. mineralisation type
          6. continuity of grade;
          7. anomalies (inconsistent or contradictory data);
          8. statistics results eg variance, CV, nugget effect, short range
          structures, neighbourhood test results;
          9. estimation method;
          10. run flag;
          11. kriging variance (relatively high or low within the strand);
          12. tonnage / metre intersected in cases of sparse drilling;
          13. confidence intervals (via gaussian transforms);
          14. conditional simulation results.

          And there are probably more. Our classification is based on a (subjective?)
          assessment of the first 10 above. And we also recognise that one poor
          measure can destroy confidence eg no amount of good interpretation or
          statistics can make up for poor sampling - the deposit will have a low
          classification.

          Hope this makes sense - and as Clint Eastwood says 'works for me!'

          Jim
          ph 08 9327 2263
          fax 08 9327 2294


          -----Original Message-----
          From: Isobel Clark [mailto:drisobelclark@...]
          Sent: Wednesday, 27, February 2002 1:40
          To: Bill Shaw
          Cc: ai-geostats@...
          Subject: Re: AI-GEOSTATS: What are appropriate measures of reliability
          for Mineral Resource and Ore Reserve estimates?


          Bill

          There are several presentations being given this week
          at the APCOM in Phoenix. You might want to look at the
          proceedings and check through those.

          One of the papers I have heard so far compares three
          criteria and shows how they affect a particular case
          study. The three are a weighted distance criterion, a
          kriging variance and an information measure. In this
          particular case, kriging variance came out bottom.

          Pity, because that's what I use!
          Isobel

          Isobel Clark
          http://uk.geocities.com/drisobelclark

          __________________________________________________
          Do You Yahoo!?
          Everything you'll ever need on one web page
          from News and Sport to Email and Music Charts
          http://uk.my.yahoo.com

          --
          * To post a message to the list, send it to ai-geostats@...
          * As a general service to the users, please remember to post a summary of
          any useful responses to your questions.
          * To unsubscribe, send an email to majordomo@... with no subject and
          "unsubscribe ai-geostats" followed by "end" on the next line in the message
          body. DO NOT SEND Subscribe/Unsubscribe requests to the list
          * Support to the list is provided at http://www.ai-geostats.org


          ----------------------------------------------
          This email (including all attachments) may be confidential. If you are not
          the intended recipient, you must not use the information contained in it.
          If you have received this email in error, please delete the email and notify
          the sender.

          --
          * To post a message to the list, send it to ai-geostats@...
          * As a general service to the users, please remember to post a summary of any useful responses to your questions.
          * To unsubscribe, send an email to majordomo@... with no subject and "unsubscribe ai-geostats" followed by "end" on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list
          * Support to the list is provided at http://www.ai-geostats.org
        • Isobel Clark
          Jim I totally agree with your principles. The main difference, in operation not philosophy, is that most (good?) practitioners include all of the above
          Message 4 of 7 , Feb 26, 2002
          • 0 Attachment
            Jim

            I totally agree with your principles. The main
            difference, in operation not philosophy, is that most
            (good?) practitioners include 'all of the above'
            before getting to a classification stage.

            For example, the selection of which blocks to estimate
            in the first place is done on the basis of all
            existing geological knowledge plus assumptions about
            continuity etc. So, a block is not even estimated
            unless it is pretty adequately known anyway. The use
            of a single criterion to distinguish the line between
            measured and indicated resources is almost a formality
            and probably depends more on the target audience than
            any other motivation.

            Isobel




            __________________________________________________
            Do You Yahoo!?
            Everything you'll ever need on one web page
            from News and Sport to Email and Music Charts
            http://uk.my.yahoo.com

            --
            * To post a message to the list, send it to ai-geostats@...
            * As a general service to the users, please remember to post a summary of any useful responses to your questions.
            * To unsubscribe, send an email to majordomo@... with no subject and "unsubscribe ai-geostats" followed by "end" on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list
            * Support to the list is provided at http://www.ai-geostats.org
          • Farquhar, James (HI)
            Isobel While your comment a block is not even estimated unless it is pretty adequately known is valid, practised on many sites, and useful eg used to justify
            Message 5 of 7 , Feb 26, 2002
            • 0 Attachment
              Isobel

              While your comment "a block is not even estimated unless it is pretty
              adequately known" is valid, practised on many sites, and useful eg used to
              justify more drilling, it is not always practical. Some software such as
              Whittle cannot cope with unestimated blocks, so we estimate a grade for them
              (eg by using larger searches), but they are classified as lower confidence.
              If not estimated, there is still a need to have some statement of this
              resource.

              Related to this, is what I term the 'salt& pepper' problem. If blocks are
              assigned a classification based on individual parameters, it can result in
              a mosaic of Inferred, Indicated and Measured blocks. While this may be
              absolutely, mathematically, correct, it too is impractical (for us anyway).
              Hence our strategy of making a judgement based on many parameters, and
              applying the classification to a 'large' contiguous area.

              I am very pleased to see this topic raised as some guidelines are long
              overdue.

              Jim
              ph 08 9327 2263
              fax 08 9327 2294


              -----Original Message-----
              From: Isobel Clark [mailto:drisobelclark@...]
              Sent: Wednesday, 27, February 2002 12:32
              To: Farquhar, James (HI)
              Cc: ai-geostats@...; Bill Shaw
              Subject: RE: AI-GEOSTATS: What are appropriate measures of reliability
              for Mineral Resource and Ore Reserve estimates?


              Jim

              I totally agree with your principles. The main
              difference, in operation not philosophy, is that most
              (good?) practitioners include 'all of the above'
              before getting to a classification stage.

              For example, the selection of which blocks to estimate
              in the first place is done on the basis of all
              existing geological knowledge plus assumptions about
              continuity etc. So, a block is not even estimated
              unless it is pretty adequately known anyway. The use
              of a single criterion to distinguish the line between
              measured and indicated resources is almost a formality
              and probably depends more on the target audience than
              any other motivation.

              Isobel




              __________________________________________________
              Do You Yahoo!?
              Everything you'll ever need on one web page
              from News and Sport to Email and Music Charts
              http://uk.my.yahoo.com


              ----------------------------------------------
              This email (including all attachments) may be confidential. If you are not
              the intended recipient, you must not use the information contained in it.
              If you have received this email in error, please delete the email and notify
              the sender.

              --
              * To post a message to the list, send it to ai-geostats@...
              * As a general service to the users, please remember to post a summary of any useful responses to your questions.
              * To unsubscribe, send an email to majordomo@... with no subject and "unsubscribe ai-geostats" followed by "end" on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list
              * Support to the list is provided at http://www.ai-geostats.org
            • Isobel Clark
              Jim, I may be intolerant in this, but I do not accept software limitations as justification for making policy decisions. There should be no reason why you
              Message 6 of 7 , Feb 26, 2002
              • 0 Attachment
                Jim,

                I may be intolerant in this, but I do not accept
                software limitations as justification for making
                policy decisions. There should be no reason why you
                cannot flag the 'dummy' blocks as inferred.

                > Related to this, is what I term the 'salt& pepper'
                > problem. If blocks are
                > assigned a classification based on individual
                > parameters, it can result in
                > a mosaic of Inferred, Indicated and Measured blocks.
                It is possible to classify your resources on the basis
                of larger units than those used to (say) optimise an
                open pit. Anglo-American, for example, generally use
                the forst two or three year's production as a single
                unit and classify that as proven/probable etc
                reserves.

                A mosaic of differently classified blocks is, surely,
                an indication that the overall sampling is inadequate
                to characterise the deposit.

                > I am very pleased to see this topic raised as some
                > guidelines are long overdue.
                I'm sure we all welcome this extension to the current
                debate. Perhaps it should be pointed out that Steve
                Henley has been running a 'chat' facility for almost
                two years on the same topic and getting little
                response.

                Isobel

                __________________________________________________
                Do You Yahoo!?
                Everything you'll ever need on one web page
                from News and Sport to Email and Music Charts
                http://uk.my.yahoo.com

                --
                * To post a message to the list, send it to ai-geostats@...
                * As a general service to the users, please remember to post a summary of any useful responses to your questions.
                * To unsubscribe, send an email to majordomo@... with no subject and "unsubscribe ai-geostats" followed by "end" on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list
                * Support to the list is provided at http://www.ai-geostats.org
              • Marcel Vallée
                Bill I was not yet aware of the JORC call for submissions on this subject. Thanks for the information. See my responses in between the key paragraphs of your
                Message 7 of 7 , Feb 27, 2002
                • 0 Attachment
                  Bill

                  I was not yet aware of the JORC call for submissions on this
                  subject. Thanks for the information.

                  See my responses in between the key paragraphs of your
                  text.

                  I have published a considerable amount of material on
                  these subjects, and the following comments only scratch
                  the top of the subject. Some of that material, I could email
                  to you but the printed material would be more convincing.
                  So I need your address.

                  Good luck in this endavour

                  Marcel Vallée Eng., Geo.

                  Géoconseil Marcel Vallée Inc.
                  706 Routhier Ave
                  Sainte-Foy, Québec G1X 3J9
                  Canada
                  Tel: (1) 418 652 3497
                  Fax: (1) 418 652 9148
                  Email: vallee.marcel@...

                  =======================================

                  02-02-25 20:07:59, "Bill Shaw" <wshaw@...>
                  wrote:
                  >
                  > From: "Bill Shaw" <wshaw@...>
                  >
                  > To: <ai-geostats@...>
                  > Subject: AI-GEOSTATS: What are appropriate
                  > measures of reliability for Mineral Resource and Ore
                  > Reserve estimates?
                  > Date: Tue, 26 Feb 2002 09:07:59 +0800
                  > Organization:Golder Associates
                  >
                  > Greetings
                  >
                  > You may be aware that in Australia the Joint Ore Reserve
                  > Committee (JORC) s called for submissions regarding
                  > the next revision of the JORC Code. The current JORC
                  > Code effective September 1999) can be downloaded
                  > from www.jorc.org and the call for submissions is also on
                  > that site.
                  >
                  > I am involved in the teams looking at the definitions for
                  > Mineral Resources and for Ore Reserves.
                  >
                  > Frequently questions are raised about the expected
                  >reliability of estimates for Resources classified as
                  > Measured, Indicated or Inferred or for Ore Reserves
                  > classified as Proved or Probable. There are a number of
                  > arguments, both for and against, regarding attempting to
                  > quantify the accuracy (and precision?) of such estimates.
                  >
                  > The Joint Ore Reserve Committee is considering whether
                  > it may be timely to introduce some discussion into the
                  > Code regarding the pros and cons of such quantification
                  > of expectations.
                  >
                  > Not all such arguments for and against quantifying the
                  > reliability of estimates are technical, for example:

                  > · FOR - The primary purpose of the Code is to
                  ensure Public Reports best inform investors. It thus
                  appears important that similarly classified
                  estimates have similar ?reliability? even if such reliability
                  can only be broadly generalised, rather than quantified.
                  >
                  > · AGAINST - The JORC Code is not currently
                  prescriptive and relies on the Competent Person making
                  decisions about many issues including
                  >
                  Comment by MV: --------------------------------------------------------

                  Complete non-prescriptivity raises a basic issue. Can one
                  reasonably expect that results may be comparable
                  between two QPs or CPs or different projects under this
                  condition. My opinion is that objectives and some basic
                  requirements basic requirements should be formulated to
                  ensure a minimum amount of focussing on what is actually
                  required in a particular case, and still leave the QP / CP a
                  considerable of authonomy in the practice of his profession.


                  --------------- End of Comment ------------------------

                  > Transparency, Materiality and Competence (all defined
                  > in the Code).
                  > Quantification of ?expected reliability? presumably
                  > requires estimates of tonnes and grades to be made
                  > using a method that would allow ?errors? to be also
                  > estimated.
                  > This may make the Code more prescriptive if such errors
                  > are only provided by certain estimation methods.

                  ----------------- Comment by MV ---------------------

                  Non prescriptivity again! It this part of some sort of a mining
                  gospel or koran!

                  ------------------End of Comment ---------------------

                  > Either way, if the JORC Code discusses this issue it must
                  > get it right. For example it is not very meaningful to
                  > suggest that Measured Resources are within +/- 10%? if
                  > such a statement can never be tested or demonstrated.
                  > Thus it has been suggested that confidence limits and
                  > parcel size must also be stated.

                  --------------Comment by MV -------------------------

                  All solutions need not be complicated. In most cases it
                  helps to try and break down the problem first. After the
                  recent crop of revisions, we have an international
                  resource/reserve system but we still have the same
                  problems because, in my opinion, the people that mattered
                  wanted to have a SIMPLE and EASY TO APPLY and NON
                  PRESCRIPTIVE REPORTING system, not an inventory
                  system.

                  What sort of objectives were present when juniors were, in
                  practice given the opportunity to make early ore reserve
                  announcements that are «economically mineable» on the
                  basis of «at least a preliminary feasibility study» (Canada)
                  or , in several jurisdictions like Australia, «appropriate
                  assessments that may include feasibility studies».

                  Feasibility studies that, under "generally accepted industry
                  practice" are required to launch a mining project, have
                  been swept under the rug and are not even defined in
                  recent definitions systems. Now an issuer (some are also
                  very efficient promoters) could announce an "economically
                  mineable reserve» that may never reach production!

                  Is this contributing to adequate information to the investors?
                  To more efficient mine project development?

                  - - - End of comment - - -
                  >
                  > We would appreciate your considered opinion on this
                  > specific issue: What are appropriate measures of
                  > reliability for Resource and Reserve estimates and
                  > how they are currently used by practitioners? I will
                  > undertake to collate submissions, to circulate these and >
                  > to prepare an overview for JORC. Please
                  > circulate this email to any colleagues that you believe
                  > are interested in this issue.

                  - - - - - - Comment by MV - - - - - - - -

                  For now, I would start by saying that the present resource /
                  reserve systems, have been designed without taking into
                  accounts the fact that mineral resource and ore reserve
                  statements, are primarily INVENTORIES for work purposes
                  BEFORE they become REPORTING documents, despite
                  the fact that you've got to estimate it before you report it.
                  Neither the formulations nor the requirements for reporting
                  carry any objetive related nor quantitative requirement.

                  Here is an example how useful quantitative requirements
                  may be worked in with limited prescriptivity and without
                  significant hardship.

                  More than 25 years ago, as Chief Geologist, Mines for the
                  SOQUEM group, I made minor adaptations to the then
                  current Canadian reserves definitions (proven, probable
                  possible) to use in our operating mines and advanced
                  projects. The chief geologist at the Louvem base metal
                  mine (his name is Réal Bourassa) had read the revised
                  text and was prepared for my next visit to Val d'Or..

                  "Marcel, your definitions are OK, but we have a problem:
                  they only provide two levels for delimined reserve (proven
                  and probable) , but we have four levels of information in the
                  mine with, in addition, the possible reserve level.

                  "In the proven category, we have drill proven ore on a 7.5 m
                  section spacing and we develop for long hole open stoping
                  or shrinkage stoping - undercut at the base overcut at the
                  summit, and raises, plus drilling at 12f ft in the ends of zones
                  to check projections. Our probable category is based on
                  two drill section spacings grids, 15m and 30 m. Would it be
                  possible to adapt "your definitions" to take all this into
                  account. We did!

                  For inventory purposes, we subdivided the proven and
                  probable levels to have four categories: «proven
                  developed », «proven drilled», «probable 1 and probable
                  2». These could be regrouped for reporting, but in such
                  cases, I insisted on supplying with each global category the
                  information regarding drilling grid size and the percentage/
                  and grade for each grid dimension. We rapidly found that
                  this system made the work of the geologist easier to
                  understand for mining engineers, mine supers, managers
                  and directors, as wall as easier to manage for the
                  geologist.

                  Another of our operating mines, with much larger and
                  multiple zones, was a 50/50 joint venture. . The mine
                  geologist had more leeway, as the partner was the mine
                  operatior and, despite my advice, he decided that is was
                  not usefull to distinguish between the probable 1 and
                  probable 2 categories. After several years of exploring for
                  new zones on a 60 m spacing and a first filling next to
                  promising intersections at 30 m spacing, the area
                  manager refused to grant additional exploration /
                  delineation budgets in that sector, as he could not see from
                  the reserve statements any significant progress in the past
                  year or two. I was called in. This problem was solved fairly
                  easily by recompiling the last two or three reserve
                  statements with the probable sub-categories, to show the
                  progress in moving large tonnages from probable 2 to
                  probable 1. Being a reasonable person, he granted the
                  budgets required to carry on.

                  Whether with four or with only the two current information
                  levels, more information could be available to the investors,
                  other industry professionnals managers and mining
                  analysts that no resource/reserve information should be
                  published without giving a «comprehensive summary» (!) of
                  the sampling information network includind number of drill
                  holes and samples, and all grid spacings for each
                  category wit. Sorry, but somebody has to be prescriptive
                  sometime when information needs to be provided.

                  This does not mean to be inflexible or unadapted. One
                  question I have frequently been asked is: What is the
                  requirement for proven ore? My answer is objective based
                  and generic:

                  ««The sampling grid for proven ore required must be
                  adapted to the deposit size, mineralization grade and
                  distribution, and be adequate to support the engineering,
                  marketing and economic studies required for project
                  feasibility for a new project, for efficient mine extraction,
                  mineral processing/ metallurgy for specific mining and
                  processing methodsin an operating mine. »»

                  This somewhat prescriptive, but implementation certainly
                  requires professional experience and judgement as well as
                  team work.

                  In summary, it probably would not be worthwhile to revise
                  the JORC guidelines, if they are to be kept non-prescriptive
                  in the present fashion and if the people involved do not
                  introduce the essential concepts required.

                  These include i) a more adequate model of the
                  development and mining process is required, ii) both
                  global and phase objectives. From these objectives, can
                  be developed explicit but fairly high level requirements that
                  will help the QPs and CPs to carry out their professional
                  responsibilities more efficiently.

                  ------- End of Response --------------------

                  > Regards,


                  > Bill Shaw
                  > wshaw@...
                  >





                  --
                  * To post a message to the list, send it to ai-geostats@...
                  * As a general service to the users, please remember to post a summary of any useful responses to your questions.
                  * To unsubscribe, send an email to majordomo@... with no subject and "unsubscribe ai-geostats" followed by "end" on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list
                  * Support to the list is provided at http://www.ai-geostats.org
                Your message has been successfully submitted and would be delivered to recipients shortly.