CIM MineSpace 2001
Olivier Tavchandjian, Lawrence Cochrane,
Abstract Resource modeling is the basis of any economic appraisal of a mineral project and includes a number of steps from data acquisition and validation to resource reporting and classification.

At each step in the modeling process, it is necessary to define the specific objectives, the methodology proposed to address those objectives and to establish a set of checks and validation tools to assess the effectiveness of the proposed methodology in addressing these issues.

The data used has to be both reliable and relevant for the purpose of the modeling. QA/QC for database validation usually involves the validation of sample location: collar location, topographic model and surveying of drill holes. Validation is also performed on assaying information: check for sampling bias and accuracy for core splits, pulp duplicates and standards, handling of calculated values, explicit/implicit absent data and data entry errors.

The second step requires establishing the geological domains, which are affected by the mineralization process. This step involves a geometric interpretation of all relevant geological features which have influenced the spatial distribution of the mineralization including dykes, faults and alterations. In addition, a geological reference system rather than a standard Cartesian system is established to measure distances between samples.

Once the mineralized domains are established and the database has been validated, statistical and spatial description of the mineralized samples are performed in order to establish the histogram and the variogram of all the variables of interest and their correlation. Results are also used to calibrate the global resource estimate within the mineralization domain and to establish the searching strategy for grade interpolation.

The mineral envelope is filled with tetrahedral cells representing mining blocks. The average grade of these blocks is estimated using kriging techniques in order to eliminate global bias and to minimize the local errors in the block grade estimate. Visual checks are made to validate the spatial trends in grade distribution while global bias and smoothing effect are assessed from the declustered sample statistics. When required, smoothing correction may be applied using change of support techniques (i.e., lognormal correction).

Conditional simulations are used to perform sensitivity analysis and the simulation models are used if severe smoothing of spatial variability makes the interpolated model inappropriate for mine planning.

The keys to successful resource estimation are an integrated team of geologists, geophysicists, geostatisticians and mine engineers using the best available technology and reviews by outside auditors. The resource team is also responsible for performing a QA/QC program in order to measure the effectiveness of the methods in meeting the specific objectives at each step of the modeling process.
Keywords: Resource estimation, Quality control, Quality Assessment
Full Access to Technical Paper
PDF version for $20.00
Other papers from CIM MineSpace 2001