1. Sampling for Mineral Resource definition A pragmatic approach H F J Theart, SRK Consulting (South Africa) Pty Ltd

2. Some definitions - Mineral Resource The SAMREC Code (2007): A Mineral Resource is a concentration or occurrence of material of economic interest in or on the earths crust in such form, quality and quantity that there are reasonable and realistic prospects for eventual economic extraction. The location, quantity, grade, continuity and other geological characteristics of a Mineral Resource are known, or estimated from specific geological evidence, sampling and knowledge interpreted from an appropriately constrained and portrayed geological model. Mineral Resources are subdivided, and must be so reported, in order of increasing confidence in respect of geoscientific evidence, into Inferred, Indicated or Measured categories. 3. Some definitions - A Sample In statistics, a sample is a set of observations drawn from a population by a defined procedure. In the Geological Profession we refer to the observation as a sample and to a statistically defined sample as a set of samples. With regard the Resource definition the geographical location of the sample is very important. (spatial statistics) It is therefore assumed that the sample is representative of the properties of the Mineral Resource at the locality in the resource from where it was collected and that the set of samples is representative of the entire Mineral Resource or that part of it, as clearly defined in the description of the Mineral Resource 4. Description of a Mineral Resource In practice a Mineral Resource is described in terms of its: Quality Volume Relative Density From this the resource can be expressed in terms of its tonnage and grade. 5. QA/QC Quality Assurance and Quality Control in Mineral Resource Definition The reporting codes only relate QA/QC to the quality of the chemical analyses of the samples Should be a feature of all aspects considered in defining the resource and the collection of the samples. All measurements for Resources and reserves. This include amongst others: The geographical location (survey) Measurement (trenches, drill holes, channel samples) The geological context (logging) Measurements of density The sample preparation The sample analysis Database Management Interpretation of the geological model 6. The geographical location (survey) Appropriate accuracy: Definition of the elevation especially for near surface deposits. The DTM. Modern advances Satellite images; and Lidar surveys; Ground control; advanced GPS technology Down hole surveys (the bearing, dip, and distance): North seeking (magnetic) magnetic declination Gyroscopic Closing the loop (In and out surveys) Conversion to appropriate geographic reference system. 7. Measurement Unlike grab samples, the samples used in resource definition has to be representative of a specific width over which it is collected. What is considered here include: Trenches, pits, drill holes, channel samples How representative the sample is over the relevant sampling width? In drilling: core or chip recovery remedies Channel and trench sampling (not bulk sampling) remedies 8. MeasurementMeasurement Even under difficult conditions the sampling should always be representative of the material sampled 9. Core recovery - Diamond drilling How is core recovery defined Also in terms of drilling contract Example Wits Gold Carbon seam related to highest grades Shear-zone hosted gold Effect of rock fabric Appropriate drilling method 10. MeasurementAppropriate drilling method By selecting the correct drilling method much improved core recoveries can be achieved. In situations such as this a double or triple core barrel design would have given much better core recovery. 11. MeasurementSample recovery Cutting channel samples with a diamond saw can give much better results but even here control measures are required 12. MeasurementSample recovery Sampling a rock dump is extremely difficult unless reverted to bulk sampling Note the pen used for scale in both cases Samples were rejected in audit as they were not representative of the material sampled. 13. The geological context The opinion of the Competent Person Is the sampling method appropriate for the style of mineralization The need for an auditable relational (digital) database An independent permanent record The core itself Core photographs Chip boards I do not believe in remote core logging The need to store sample pulps under appropriate conditions for repeat analysis 14. Is the sampling method appropriate? Vertical drill holes were used to sample the deposit. 15. The Permanent Record 16. Measurements of Density Consider using the correct terminology and not Specific Gravity. In situ dry bulk density is used to convert the resource volume to resource tonnage. Density measurements should reflect an auditable record of QA/QC as should any other resource measurement. It is good practice to confirm the measurement by at least one alternative method. 17. Dry Bulk Density Some of the alternative measurements include: The simple volume and mass method Calliper measurement, ore recovery Water displacement method Difficulty to measure displaced volume Archimedes Principle (special care with porous or water absorbing materials) The gas pycnometer method Small sample (Is it representative) Porous rocks The Stoichiometric method For this the chemical composition and mineralogy should be known 18. Problem with Resource Statement of an Iron ore deposit The company states in its Mineral Resource Statement (JORC compliant) that the in situ bulk dry density is 2.80. And this is used for the tonnage conversion. A simple stoichiometric calculation indicates that the Relative Density should be in the order of 4.94. Clearly there is a problem and one should carefully consider the companys defence that the lower density is a conservative estimate. My client was the government who can be negatively influenced by such an action. Example 19. The core 20. Overestimation that may be directly related to RD and sample quality Nevsun Resources Ltd. (NSU) on 7 February announced its expected 2012 outlook for planned gold production is 190,000 to 210,000 ounces from the Bisha mine in Eritrea. This is about half of what Nevsun was previously expecting to produce in 2012. They concluded that the resource estimate used for mine planning over-estimated gold in portions of the Bisha Main Oxide Mineralization. It goes on to add that its new resource estimate includes approximately 1.3 million tonnes @ 5.1 g/t Au of the remaining 1.6 Mt of Bisha Main Oxide. This compares with 2.5 Mt that was previously expected. Possible causes: Core recovery Inappropriate drilling method - reverse circulation drilling Gas pycnometer density determinations Example 21. Sample preparation Sample preparation is often the cause of the biggest QA/QC problems and this include: Splitting error Ni sulphide example Cross contamination Vein gold example External contamination Vertical spindle mills Consider the use of blank samples very carefully, especially the point at which they are introduced in the process 22. The chemical analysis of samples Most advanced in terms of QA/QC Laboratory accreditation ISO 17025 Umpire laboratory analyses Blanks, duplicates and standards (in house as well as Certified Reference Materials). More important questions are: Is the sample dissolution appropriate Is the analytical method appropriate 23. Database Management The most important aspect that I want to address here is the need to reduce human error as far as possible. Format the database and the laboratory report so that there is a seamless import of the data and merging with various other types of data to avoid input errors and column swaps. Ensure that there is adequate access control and database security to prevent unauthorised access. Ensure that access to the data base is limited, recorded and that there is sign-off by responsible persons. Build automatic QC processing into the database so that problems can be identified and appropriate measures taken as soon as possible. Design to facilitate easy import and query procedures 24. Interpretation of the geological model Sampling design must be appropriate for the geological model Sample density Definition of the geological model is normally based on: 3-D distribution of information. Geological structures, and intrusions that will affect the distribution of the ore zone inferring continuity across such features may lead to mistakes in the estimation of grade. Delineation of the ore zone often require assumptions with regard cut-off grades, the definition of geological losses, and the certainty with which the geographic position of sources of information. Sample information must be distributed in a manner that demonstrates that they represent the overall characteristics of the resource or zone within the resource. It should be clear what is done with regard missing data. Example massive sulphides in a Ni-sulphide ore body