4
Indicator Kriging process, was created by accumulating
the 9 simulations of each block into a histogram for each
block. For a given volume, the IK model shows the same
histogram as the sum of the 9 CS histograms, but the IK
model loses the bench to bench correlation of individual
simulations. The “OK” model, approximating an Ordinary
Kriging process, was created by averaging the 9 simulated
grades for each block the OK model shows significantly less
variability (i.e., less selectivity) than the IK or CS models.
Base Case Mine Plan from Deterministic Optimization
A base case mine plan (Case C: mill only with stockpiling)
was built using deterministic optimization with IK model
inputs. Figure 2 shows dashboard graphs summarizing key
aspects of the base case mine plan. Tons mined are con-
strained by vertical advance in year 1 and by total mining
capacity in subsequent years. Tons milled are constrained
by mill capacity in all years with the mill vs stockpile cut-off
controlled by the grade-tonnage distribution and remain-
ing material above 0.02 oz/ton cut-off stockpiled for pro-
cessing after mining is completed.
Base Case Stochastic Evaluation
Figure 3 shows key aspects of the stochastic evaluation
of the base case mine plan. For the base case mine plan
defined in Figure 2, nine conditional simulations yield nine
grade/tonnage distributions for each time period’s mining
(Figure 3, upper left). For each mined grade/tonnage distri-
bution, increments are assigned to mill/stockpile/waste to
balance tons milled with mill capacity. For each realization,
slightly different stockpile flows (middle left) and cut-off
grades (middle right) are selected to adapt to each realiza-
tion’s simulated grade/tonnage data. The orebody uncer-
tainties lead to uncertainties in gold production (lower
right) and cash flows (lower left).
Base Case Risk Profiles
Stochastic evaluation results can be shown for any variable
of interest in the mine plan. Figure 4 shows the uncertainty
of grades, tons and contained metal mined above a 0.02
oz/ton cut-off. Note how above-average grades above cut-
off tend to correlate with above-average tons above cut-off.
Also note how any given simulation realization tends to
stay above or below the average line for several consecu-
tive years. These are common spatial correlation patterns
for gold deposits and demonstrate why mine call factors
can be useful mine planning tools for projecting past model
reconciliation performance into future time periods.
Figure 5 summarizes the contained metal mined simu-
lations into a box plot view that shows the 25th, 50th and
75th percentiles as two rectangles along with lines (also
called whiskers) extending to min and max values and a
diamond representing the average of all the simulations.
Figure 6 shows how the metal mined risk profile, when
combined with mill recovery, stockpile flows, and mill ton-
nage constraints, translates to a metal produced risk profile.
The yearly risk profile in Figure 6 reflects the net
impact of drillhole spacing/drilling density and localized
grade uncertainty partially mitigated by the impacts of
multiple independent mining faces and low uncertainty for
Figure 2. Base case mine plan optimization dashboard
Indicator Kriging process, was created by accumulating
the 9 simulations of each block into a histogram for each
block. For a given volume, the IK model shows the same
histogram as the sum of the 9 CS histograms, but the IK
model loses the bench to bench correlation of individual
simulations. The “OK” model, approximating an Ordinary
Kriging process, was created by averaging the 9 simulated
grades for each block the OK model shows significantly less
variability (i.e., less selectivity) than the IK or CS models.
Base Case Mine Plan from Deterministic Optimization
A base case mine plan (Case C: mill only with stockpiling)
was built using deterministic optimization with IK model
inputs. Figure 2 shows dashboard graphs summarizing key
aspects of the base case mine plan. Tons mined are con-
strained by vertical advance in year 1 and by total mining
capacity in subsequent years. Tons milled are constrained
by mill capacity in all years with the mill vs stockpile cut-off
controlled by the grade-tonnage distribution and remain-
ing material above 0.02 oz/ton cut-off stockpiled for pro-
cessing after mining is completed.
Base Case Stochastic Evaluation
Figure 3 shows key aspects of the stochastic evaluation
of the base case mine plan. For the base case mine plan
defined in Figure 2, nine conditional simulations yield nine
grade/tonnage distributions for each time period’s mining
(Figure 3, upper left). For each mined grade/tonnage distri-
bution, increments are assigned to mill/stockpile/waste to
balance tons milled with mill capacity. For each realization,
slightly different stockpile flows (middle left) and cut-off
grades (middle right) are selected to adapt to each realiza-
tion’s simulated grade/tonnage data. The orebody uncer-
tainties lead to uncertainties in gold production (lower
right) and cash flows (lower left).
Base Case Risk Profiles
Stochastic evaluation results can be shown for any variable
of interest in the mine plan. Figure 4 shows the uncertainty
of grades, tons and contained metal mined above a 0.02
oz/ton cut-off. Note how above-average grades above cut-
off tend to correlate with above-average tons above cut-off.
Also note how any given simulation realization tends to
stay above or below the average line for several consecu-
tive years. These are common spatial correlation patterns
for gold deposits and demonstrate why mine call factors
can be useful mine planning tools for projecting past model
reconciliation performance into future time periods.
Figure 5 summarizes the contained metal mined simu-
lations into a box plot view that shows the 25th, 50th and
75th percentiles as two rectangles along with lines (also
called whiskers) extending to min and max values and a
diamond representing the average of all the simulations.
Figure 6 shows how the metal mined risk profile, when
combined with mill recovery, stockpile flows, and mill ton-
nage constraints, translates to a metal produced risk profile.
The yearly risk profile in Figure 6 reflects the net
impact of drillhole spacing/drilling density and localized
grade uncertainty partially mitigated by the impacts of
multiple independent mining faces and low uncertainty for
Figure 2. Base case mine plan optimization dashboard