Geoscience Reference
In-Depth Information
zones within the deposit correspond to tonnages and grades
planned to be extracted according to an existing mine sched-
ule. In this sense, the set of 10 simulated values for each area
is used to provide a risk assessment on the existing mine
schedule and predicted cash flows, as well as comparing it
with the more traditional resource classification scheme.
Only 4 of the 17 areas are discussed here, 2 open pit areas
and 2 underground areas corresponding to a medium-term
planning horizon. The open pit areas are the D4 and Mining
Phase 7 within the Lince pit. The underground areas shown
are the A1 and the D1/D2 combined area. For convenience,
each of these areas was represented in the computer with a
three-dimensional solid, such that these solids can be used
to select the blocks of interest and calculate tonnages and
grades.
There are several options available to both process the in-
formation and perform the mining risk analysis and to pres-
ent the model of uncertainty. The following information is
presented here:
• Average of the area according to the Resource Block
Model (“Ave. Res. Model” on the graphs).
• Average of the area according to the simulated values
(“Ave. Sims” on the graphs).
• Lower Probability Limit, defined as the 15th percentile
of the distribution from the simulation model (“Lower
Limit” on the graphs).
• Upper Probability Limit, defined as the 85th percentile of
the distribution of possible values for each block (“Upper
Limit” on the graphs).
The results are presented for four different cutoff grades: 0 %
TCu (or global), 0.5 % TCu, 1.0 % TCu, and 1.2 % TCu. The
results are expressed in % TCu, and the Upper and Lower
Probability Limits are such that there is a 70 % probability
that the true value is within those limits. In addition, for
each area, the results are presented by Resource Classifica-
tion categories (measured and indicated), in addition to total
resources (which includes inferred as well as the previous
two). Figures
14.53
-
14.56
present these results.
The results were obtained using volume weighing for
each area; in other words, the metal content for each cutoff
and classification category were first obtained, and from that
the grades presented in Figs.
14.53
-
14.56
were derived.
the averages at a 0 % cutoff are more similar. The issue
of recoverable reserves is best addressed through condi-
tional simulations, not a more traditional change of sup-
port model applied to the MIK model (Chap. 7, and also
Journel and Kyriakidis
2004
; Rossi and Alvarado
1998
).
2. The simulation model results in probability intervals that
are not symmetric with respect to the expected value.
There is no reason why the probability of error on one
side of the expected value has to be identical to the prob-
ability of error of the opposite side.
3. It is possible that the expected value (according to the
Resource Model) falls outside the probability limits
(
P
85
-
P
15
) defined; this can happen because the simulation
model is obtained independently of the estimation model
(even if applying the same random function), and is more
likely when cutoff grades (conditional statistics) are con-
sidered.
4. The probability intervals are different for each cutoff
grade analyzed. It is also different from measured to in-
dicated to inferred resources. In general, higher cutoff
grades result in wider probability intervals (higher uncer-
tainty and risk, as expected), and the same can be said for
the difference between measured, indicated, and inferred
resources.
5. The measured, indicated, and inferred classes are not very
useful when analyzing uncertainty and risk within local
areas. For example, the measured category in one area
will have a different uncertainty than the same measured
category in a different area. The reason is that the resource
classification scheme is usually developed on a global
basis, and at best only appropriate for long-term risk as-
sessments. The conditional simulation model shows that
it is not appropriate to use the resource classification
schemes for local risk assessment. That is, a certain block
can be classified as measured within the long-term con-
text, but it may not be even indicated when shorter term
production periods are considered.
6. The resource classification categories for different areas
show significant variability when described in terms
of probabilities. For example, the measured resources
for Phase 7 Expansion (for the 0.5 % TCu cutoff) show
that the 70 % probability interval is within + 16 %/− 8 %
(24 % total width) of the resource model grade, while
the same “measured” resources for Area D1/D2 is with-
in + 8 %/− 20 % of the resource model grade. What is
called measured in one area with a given probability in-
terval may have significantly different probability inter-
vals in another area, but still be called “measured”. This
is to be expected, due to local geologic differences, ad-
ditional complexities of the mineralization controls, and
local differences in drill hole coverage.
Additional area-specific comments are:
14.5.6
Results
In addition to specific conclusions for each Area described
below, the following are some of the more important general
observations and conclusions:
1. The average grade of the simulations is somewhat differ-
ent than the average grade from the resource model for
most areas. This is a consequence of the differences in the
internal dilution incorporated into each model; note how
