Training for Glencore on Geostatistics and Resource Modeling

Classification
May 2022
Context
• Classification is required for public disclosure of
resources and reserves for listed companies
• What – bin the estimates
• When – before any public disclosure
• Who – the qualified person with all relevant technical
support
• How – data spacing with geological and technical support

• Many widely varying classification methods used:

https://www.crirsco.com/docs/revised_standard_definitions2011_v2.pdf
• Search neighbourhoods, data spacing, kriging
performance; others not described

• Each technique will give a different result

• Manage with expert judgement, best local practice, reconciliation
of previous classification results, score, and peer review

2
Learning Outcomes
• List the paradigms of numerical geological modeling
• Describe uncertainty as a consequence of heterogeneity and sparse sampling
• Name the core techniques used in geostatistical modeling
• Discuss applications of geostatistics in resource management

3
Definition, Purpose and Guidelines

4
Classification
• Put estimates of ore (tonnes, grade, metal) into different bins for public disclosure
• Permit SECs and investors to judge value in a fair manner

5
Purpose of Classification
• SECs require some standards for disclosure to protect investors

• Classification required to inform public on mineral resources and reserves that a company expects to be
able to economically extract

• Emphasis is on reasonable and transparent disclosure to limit liability and ensure even dissemination of
information

• Considers both ore grades as well as all relevant technical, economical, environmental, and social
modifying factors

• Most major mining districts define requirements through their mining association although the US is a
notable exception with Industry Guide 7

6
Resource Definitions
• Measured Mineral Resource:
Mineralization or other natural material of economic interest may be classified as a Measured Mineral Resource by the
Qualified Person when the nature, quality, quantity and distribution of data are such that the tonnage and grade of the
mineralization can be estimated to within close limits and that variation from the estimate would not significantly affect
potential economic viability. This category requires a high level of confidence in, and understanding of, the geology and
controls of the mineral deposit.

• Indicated Mineral Resource:

Mineralization may be classified as an Indicated Mineral Resource by the Qualified Person when the nature, quality,
quantity and distribution of data are such as to allow confident interpretation of the geological framework and to
reasonably assume the continuity of mineralization. The Qualified Person must recognize the importance of the
Indicated Mineral Resource category to the advancement of the feasibility of the project. An Indicated Mineral Resource
estimate is of sufficient quality to support a Preliminary Feasibility Study which can serve as the basis for major
development decisions.

7
Reserve Definitions
• A Mineral Reserve is the economically mineable part of a Measured or Indicated Mineral Resource
demonstrated by at least a Preliminary Feasibility Study. This Study must include adequate information
on mining, processing, metallurgical, economic and other relevant factors that demonstrate, at the time
of reporting, that economic extraction can be justified. A Mineral Reserve includes diluting materials and
allowances for losses that may occur when the material is mined.
• A ‘Proven Mineral Reserve’ is the economically mineable part of a Measured Mineral Resource demonstrated by at
least a Preliminary Feasibility Study. This Study must include adequate information on mining, processing,
metallurgical, economic, and other relevant factors that demonstrate, at the time of reporting, that economic
extraction is justified.
• A ‘Probable Mineral Reserve’ is the economically mineable part of an Indicated, and in some circumstances a
Measured Mineral Resource demonstrated by at least a Preliminary Feasibility Study. This Study must include
adequate information on mining, processing, metallurgical, economic, and other relevant factors that
demonstrate, at the time of reporting, that economic extraction can be justified.

8
Guidelines for Classification
• CRIRSCO (Committee for Mineral Reserves International Reporting Standards) is composed of many
major mining associations around the world
• Defines a template which is generally accepted

http://www.crirsco.com/images/crirsco_members_map_02.jpg
9
NI 43-101
• Instrument accepted by Alberta, British Columbia, Toronto Stock Exchange… for mineral resources and
reserves disclosure

• Defines resources, reserves, and requirements on the authors and contents of technical reports

• Competent persons require 5 years of experience

• Largely compatible with other codes:

• PERC (Pan-European Reserves and Resources Reporting Committee)

• SAMREC (South African Mineral Resource Committee)
• JORC (Joint Ore Reserves Committee of the Australasian Institute of Mining and Metallurgy)

10
Principles

11
Principles of Classification
• Principles of classification defined in all guidelines and as part of the CRIRSCO template:
• Materiality – present facts (such as assays, sampling protocols) and evidence supporting these such
as quality assurance and quality control procedures met and followed

• Transparency – presentation of classification methodology should be unambiguous

• Competency – required qualified/competent person with experience in the field to sign on reports

• Modifying Factors – considers all relevant technical and socio-economic factors when defining
reserves

• All completed within the context of “demonstrated geological continuity”

12
Materiality and Transparency
• Analysis based on assays and geologic interpretations with quality assurance and quality control
procedures followed for assays to ensure reliability

• Geologic interpretations must establish geologic continuity for the level of classification

• Drilling must support geologic continuity

• Clearly communicate resource model construction process and classification criteria

• Minimize room for mischief and focus on openness using clear language

13
Competency
• Standards vary, but typically include:
• Member of professional body with an enforceable code of ethics such as a Professional Geoscientist
or Professional Engineer and is in good standing

• University degree in area associated with mining (such as geology, metallurgy, mining engineering,
environmental engineering, among others)

• Five years experience relevant to the type of mineralization, deposit, and mining activity

• Qualified or competent person liable for many aspects of disclosure

• Professional association must admit based on 1) academic qualifications, 2) experience, and 3) ethical
requirements and must require or encourage continuing professional development

• Professional association must be able to revoke membership

14
Modifying Factors
• Modifying factors may be any factors which impact the potential economic viability of the project
including, but not limited to:
• Mine-ability
• Processing and metallurgical constraints
• Environmental factors
• Government
• Marketing and sale of commodity
• Economic factors including transportation and sales costs
• Social factors
• Land, lease and royalty considerations

15
Codification of Projects

https://www.unece.org/fileadmin/DAM/energy/se/pp/unfc/ws_UNFC_MexicoCity_Sept2012/Session6_Edmundo_Tulcanaza.pdf
16
Methods of Classification

17
Historical Approaches
• Expert assessment of geological continuity
• Distance from a sample(s)
• Sample density in the vicinity of a block
• Geometric configuration of data available to estimate a block
• Kriging variance or relative kriging variance

18
Distribution of Approaches
• 2013 study of NI-43101 approaches from Silva (2015)
• Note the location and commodity types

19
Distribution of Approaches (2)
• Methods used for resource estimation
• Methods used for resource classification
• Neighbourhood restrictions (NR), such as a search ellipse
• Drill hole spacing (DHS)
• Kriging variance (KV)
• Regression slope (RS) from kriging
• Kriging efficiency (KE)
• Drill hole spacing and neighbourhood restrictions are the most popular

20
Search and Data Spacing Approaches

21
Search Neighbourhoods
• Referred to as neighborhood restrictions in Silva (2015)

• Search restrictions common:

• Min number of drill holes
• Max number of samples per drill hole
• Min and max number of samples
• Max number of samples per octant
• Anisotropic radius (typically follows variogram)
• Covariance sorted search
• High grade restrictions

• Target search for the desired model goal

• Artifact free visualization
• Interim estimate for resource model
• Grade control

http://pro.arcgis.com/en/pro-app/help/analysis/geostatistical-analyst/search-neighborhoods.htm
22
More Comments on Search
• Block size chosen as SMU size is convenient
• Potentially larger to account for additional dilution and information effect

• Discretization controlled by composite length in

one direction and enough (4x4) in other

• Set search ellipsoid quite large and let kriging

and number of data sort out the weights
• Maximum per drill hole
• Maximum per octant

• Consider clipping by depth, convex hull of the

data, maximum data spacing…

23
Multiple Pass Kriging Searches
• Defining data spacing directly is challenging, so search neighbourhoods commonly used – multiple pass
kriging may be set up:
• 1st pass: krige with a minimum requirement of N samples in L anisotropic radius for measured
• 2nd pass: krige blocks unestimated from the first pass with a minimum requirement of M<N samples
in K>L anisotropic radius for indicated

• Will lead to artifacts where transition occurs which are more stronger than standard restricted kriging
artifacts

24
Variogram Range Assessment
• Alternatives are:
• Choose a fraction of the variogram range for drill spacing
• Use a fraction of the variogram range for the distance in a multipass estimation search

• Difficulty choosing the fraction and defining continuity

25
Data Spacing Based Techniques
• Calculate the density of data per unit volume
• Choice of thresholds depend on:
• Industry-standard practice in the country and geologic province
• Experience from similar deposit types
• Calibration with uncertainty quantified by geostatistical calculations
• Expert judgement of the competent or qualified person

26
Kriging Approaches

27
Kriging Performance
• QKNA attributed to Vann, Jackson and Bertoli in 2003. They strongly
advocated considering:
– Slope of regression
– Weight of the mean for SK
– Distribution of weights including negative weights
– Kriging variance

• Further augmented by:

– Kriging efficiency (Krige’s definition and statistical definition)

• Classify based on a threshold for kriging variance, slope of regression, or

kriging efficiency, or a combination of multiple factors

• All kriging variance measures are proxies for the amount of data available in
the neighbourhood of the block being classified

28
Kriging Variance
• Kriging variance minimizes the estimation variance

• Three important factors

• Variance if no data are available depends on block size
• Data in proximity of block being estimated reduce kriging variance
• Redundant data, all else being equal, increases the kriging variance
• Can threshold based on expert judgement for classification

29
Slope of Regression
• Slope can be calculated on a block by block basis in most kriging programs; for ordinary kriging:

C (V , V ) − σ OK
2
+µ
b=
C (V , V ) − σ OK
2
+ 2µ

• Slope near 1 is good; low slope is “bad” due to conditional bias, but models are not over smoothed
• Change by (1) changing block size, or (2) changing number of data

30
Kriging Efficiency
BV − KV
KEDK (%) =
• Krige’s definition: BV
• BV is expected block variance (gammabar calculations)
• KV is the kriging variance

• Measures smoothing, but not efficiency in a statistical sense

• Efficiency in Statistics
• Minimum possible variance divided by variance of the estimator
• Higher variance of estimator – low efficiency

• (Geo)Statistical Kriging Efficiency

GSKV ( u) σ GSK
2
( u)
KE ( u) =
=
KV ( u) σ K2 ( u)
• Both measures measure the departure from optimality that we are paying to reduce smoothing and
depend less on stationarity

31
Integration and Comparison of Methods

32
Integration of Multiple Methods with Scoring
• Multiple classification methods and mitigating factors can be implemented with a scoring process

• Assign every block a score for each factor:

• Geologic continuity in region
• Estimation confidence
• Data quality and integrity
• Geometallurgical factors and penalty elements
• Expert judgement

• Sum all scores weighted by their importance for classification

• Classify based on score, smoothing the results to avoid the spotted dog pattern

33
Avoid the Spotted Dog
• “Spotted Dog” classification not considered best practice
• We are more certain near drilling as indicated by the bullseye pattern, but this pattern runs contrary to
our requirement for reasonable geologic continuity

Australian Guidelines for the Estimation and Classification of Coal Resources, JORC, 2014
34
2-D Comparison
• Average drill hole spacing (ADHS)
• Cross-validation variance (CVV)

35
3-D Porphyry Comparison

36
Review of Main Points
• Classification for public disclosure of resources and reserves for listed companies
• What – bin the estimates
• When – before any public disclosure
• Who – the qualified person with all relevant technical support
• How – data spacing with geological and technical support

• Many widely varying classification methods used:

• Search neighbourhoods, data spacing, kriging performance; others not described
• Session X will discuss probabilistic classification

• Each technique will give a different result

• Manage with expert judgement, best local practice, reconciliation of previous classification results, score, and peer
review

37