199
Variography and grade estimation using ordinary kriging occur in transformed space, with
both constrained vertically by a third, to a quarter of the maximum domain thickness. This
ensures samples occurring laterally within the same position in the dilated profile gain prominence over those occurring in a different vertical position in the grade profile. This honours
the strong vertical grade and sulphide species zonation evident at Kamoa-Kakula, allowing
full vertical optimisation and material selection to be made.
A link between extensional fault architecture and localisation of orebodies is known in the
CACB. Thickness changes in modelled units were used to guide the search orientations during estimation, in a Datamine process known as ‘dynamic anisotropy’. This process allows
the search orientation to change from block to block across the deposit based on trends
incorporated into the block model, ensuring that known geological controls affect the estimate of grade.
There are significant benefits to using the 3D modelling approach:
– The vertical grade definition allows for optimisation studies during mine planning and
scheduling
– The resource model only needs to be built once as no engineering constraints are built into
the model to limit its flexibility; no additional dilution models need to be constructed
– The modelling approach can effectively model numerous mineralised horizons at different
vertical (and lateral) positions within the stratigraphic profile. Variography and estimation
thus occur within horizons that share lithological or mineralisation characteristics; estimated block grades do not use samples from cross-stratigraphic positions.
– In addition to copper grades, processing, smelter and marketing studies can rely on information in the model including copper sulphide species, arsenic, iron and sulphur grades.
The improved model definition does create some challenges:
– Reporting the resource is complex as areas occur where two or more mineralised horizons
overlap. Given the limited selectivity of underground mining techniques, isolated blocks
above cut-off were not considered economically extractable, and were excluded from the
Mineral Resource. Reporting also considers minimum mining heights.
– Processing time for the estimate is significantly longer than the 2D model, with much
larger file sizes created as outputs.
– Displaying grade trends across stratigraphic or mineralised units requires post-processing
of the model to create a 2D equivalent.
4 CONCLUSION
During the exploration phase, an emphasis is placed on understanding geological processes.
Once resource estimation and mine planning commence, the emphasis is placed on ensuring the geometry is correct. In the modelling approach developed at Kamoa-Kakula, both
the geological processes and accuracy in geometry are combined to produce a high quality
resource model.
The modelling approaches have progressively improved to include more detailed information and geological control in the grade estimates. Models have also aligned with evolving
requirements from advancing mining studies, and have added significant value to project
economics.
Models have progressed from a simplified 2D model using isotropic ID2, to a full 3D
model estimated in transformed space using ordinary kriging with dynamic anisotropy to
ensure vertical and lateral grade trends evident in the vast drillhole database are replicated in
the estimated block model grades.
Prior to the inclusion of close-spaced grade control information, resource models using
wider spaced drilling and estimated using ordinary kriging will likely be a smoother representation of the block grade distribution than encountered in reality. Many factors inherent
to the Kamoa deposit, and Kakula in particular, mitigate this risk:
Variography and grade estimation using ordinary kriging occur in transformed space, with
both constrained vertically by a third, to a quarter of the maximum domain thickness. This
ensures samples occurring laterally within the same position in the dilated profile gain prominence over those occurring in a different vertical position in the grade profile. This honours
the strong vertical grade and sulphide species zonation evident at Kamoa-Kakula, allowing
full vertical optimisation and material selection to be made.
A link between extensional fault architecture and localisation of orebodies is known in the
CACB. Thickness changes in modelled units were used to guide the search orientations during estimation, in a Datamine process known as ‘dynamic anisotropy’. This process allows
the search orientation to change from block to block across the deposit based on trends
incorporated into the block model, ensuring that known geological controls affect the estimate of grade.
There are significant benefits to using the 3D modelling approach:
– The vertical grade definition allows for optimisation studies during mine planning and
scheduling
– The resource model only needs to be built once as no engineering constraints are built into
the model to limit its flexibility; no additional dilution models need to be constructed
– The modelling approach can effectively model numerous mineralised horizons at different
vertical (and lateral) positions within the stratigraphic profile. Variography and estimation
thus occur within horizons that share lithological or mineralisation characteristics; estimated block grades do not use samples from cross-stratigraphic positions.
– In addition to copper grades, processing, smelter and marketing studies can rely on information in the model including copper sulphide species, arsenic, iron and sulphur grades.
The improved model definition does create some challenges:
– Reporting the resource is complex as areas occur where two or more mineralised horizons
overlap. Given the limited selectivity of underground mining techniques, isolated blocks
above cut-off were not considered economically extractable, and were excluded from the
Mineral Resource. Reporting also considers minimum mining heights.
– Processing time for the estimate is significantly longer than the 2D model, with much
larger file sizes created as outputs.
– Displaying grade trends across stratigraphic or mineralised units requires post-processing
of the model to create a 2D equivalent.
4 CONCLUSION
During the exploration phase, an emphasis is placed on understanding geological processes.
Once resource estimation and mine planning commence, the emphasis is placed on ensuring the geometry is correct. In the modelling approach developed at Kamoa-Kakula, both
the geological processes and accuracy in geometry are combined to produce a high quality
resource model.
The modelling approaches have progressively improved to include more detailed information and geological control in the grade estimates. Models have also aligned with evolving
requirements from advancing mining studies, and have added significant value to project
economics.
Models have progressed from a simplified 2D model using isotropic ID2, to a full 3D
model estimated in transformed space using ordinary kriging with dynamic anisotropy to
ensure vertical and lateral grade trends evident in the vast drillhole database are replicated in
the estimated block model grades.
Prior to the inclusion of close-spaced grade control information, resource models using
wider spaced drilling and estimated using ordinary kriging will likely be a smoother representation of the block grade distribution than encountered in reality. Many factors inherent
to the Kamoa deposit, and Kakula in particular, mitigate this risk: