DIMITRAKOPOULOS, R and FONSECA, M.B.

Assessing risk in grade-tonnage curves in a complex copper deposit, northern Brazil, based on an efficient
joint simulation of multiple correlated variables. Application of Computers and Operations Research in the Minerals Industries, South African Institute of
Mining and Metallurgy, 2003.

Assessing risk in grade-tonnage curves in a complex copper

deposit, northern Brazil, based on an efficient joint simulation
of multiple correlated variables
R. DIMITRAKOPOULOS* and M.B. FONSECA†
*WH Bryan Mining and Geology Research Centre, University of Queensland, Brisbane, Australia
†Companhia Vale do Rio Doce (CVRD), Brazil

Risk quantification in grade-tonnage curves is critical for capital investment in mining projects
and can be obtained through geostatistical simulations of orebodies. A practical difficulty may
arise in multi-element deposits, as the joint modelling of the related attributes using the traditional
co-simulation approaches is computationally intensive and may be impractical for use in the
industrial environment. This paper presents the construction of risk-integrating grade-tonnage
curves for a complex copper deposit in northern Brazil, by jointly simulating its key geochemical
attributes of interest: Cu, Fe and K. The joint conditional simulation of these elements is based on
Minimum/Maximum Autocorrelation Factors (MAF). MAF is an approach, based on principal
components, that spatially decorrelates the variables involved to non-correlated factors. MAF’s
spatial decorrelation at any lag distance is the main and critical difference of this approach from
the principal component approach attempted in the past. In the MAF approach, the independent
factors are individually simulated and back-transformed to the conditional simulations of the
correlated deposit attributes that reproduce the cross-correlations of the original variables.
Keywords: Joint simulation; mini/max autocorrelation factors; grade-tonnage curves.

Introduction Dimitrakopoulos8 have suggested the use of the so-termed

Capital investment in mining projects requires the
quantification, understanding and assessment of risk in
grade-tonnage curves. Geostatistical simulation
technologies1 provide an increasingly recognized tool to
model geological uncertainty and quantify geological risk
associated with grade-tonnage curves. Frequently, mineral
deposits and their geological characteristics are described
by a multitude of geochemically interrelated attributes. The
joint modelling of these attributes assists the geological
plausibility of complex orebody models as well as the
modelling of individual attributes. A key bottleneck,
however, is that the common joint simulation methods2,3 are
too computationally intensive to be of practical use in the
industrial environment, particularly when more than two
attributes are considered. Contributors to complexity
include the tedious inference and modelling of cross-
variograms, and computational inefficiency, substantially
increasing with the number of variables being co-simulated.
A practical alternative to the ‘direct’ co-simulation of
variables is the decorrelation of variables introduced by
David4. His approach, demonstrated in the joint simulation
of a uranium deposit, is based on the decorrelation of
variables using principal component analysis5–7 (PCA). The
effectiveness of this approach, in the presence of spatial
cross-correlations inherent in mineral deposits, is limited
because PCA ignores cross-correlations at distances other
than zero. To overcome the limitations of the ‘direct’ co-
simulation methods and PCA, Desbarats and
Minimum/Maximum Autocorrelation Factors, or MAF, in
the context of spatial simulation. The MAF approach may
be described as an approach, based on principal
components, that spatially decorrelates the variables
involved to non-correlated factors. The independent MAF
are individually simulated and back-transformed to the
conditional simulations of the correlated deposit attributes,
that reproduce the cross-correlations of the original
variables.
This paper presents the joint simulation of copper, iron
and potassium in an oxide copper deposit located in
northern Brazil (Figure 1) and the resulting assessment of
risk in grade-tonnage curves for copper. The geological
environment of the deposit is complex. It occurs within a
faulted sequence of Archean felsic-basic metavolcanics
hosting hydrothermal breccias and alteration haloes. Copper
is dispersed throughout the weathering profile but is also
disseminated and enriched in sheet-like saprolitic units
outside of the weathered mineralized breccia bodies. The
ability of the deposit to supply ore for a SX-EW
metallurgical process plant is a key point assessed in a pre-
feasibility study, and the joint simulation is particularly
important in assessing copper solubility controlled by iron
and potassium content.
The following sections include; a summary of the method
of joint simulation of multiple correlated variables based on
MAF; a description of the deposit and the data available;
the results of the joint simulation; the presentation of grade-
tonnage curves and risk analysis; and conclusions.

ASSESSING RISK IN GRADE-TONNAGE CURVES IN A COMPLEX COPPER DEPOSIT 373

 Joint simulation of correlated variables using
minimum/maximum autocorrelation factors
In geostatistical terminology, the attributes of a multi-
element mineral deposit are represented by a multivariate
stationary and ergodic random function. Consider a
multivariate, l dimensional, Gaussian, stationary and
ergodic spatial random function Z(x) = [Z1(x),...,Zl(x)]T. The
Minimum/Maximum Autocorrelations Factors are defined
as the l orthogonal linear combinations Yi(x) = aiT Z(x), i =
1,.., l of the original multivariate vector Z(x). MAF are
derived assuming that Z (x) is represented by a two-
structure linear model of co-regionalization6. The MAF
transformation can be rewritten as
Y ( x ) = AMAF Z ( x ) [1]
and the MAF factors are derived from
AMAF = Q2 Λ−11Q1 [2]
where the eigenvectors Q1 and eigenvalues Λ1 are obtained
from the spectral decomposition of the multivariate
covariance matrix B of Z(x) at zero lag distance. More
specifically,
Q1 BQ1T = Λ1 [3]
and Q 2 is the matrix of eigenvectors from the spectral
decomposition
2 [ T
]
Q2 M ( ∆ )Q2T = Q2  [ΓY ( ∆ )] + [ΓY ( ∆ )]  Q2T
1

[4]
where matrix ΓY(∆) is an asymmetric matrix variogram at
lag distance ∆ for the regular PCA factors Y(x)=Z(x) A,
where A = QΛ-1/2. In practice, several ∆ lag distances may
be used for values lower than the range and the resulting
eigenvectors averaged.
Given the MAF transformation above, the joint
simulation of multiple correlated variables using the MAF
approach proceeds as follows:
• Normalize the variables to be simulated
• Use MAF to generate the MAF non-correlated factors
• Produce variograms for each MAF
• Conditionally simulate each MAF using any Gaussian
simulation method
• Validate the simulation of factors
• Back-transform simulated MAF to variables and
denormalize
• Validate the final results
• Generate additional simulations, as needed.
In most cases, the reblocking of the generated realizations
to a block support model is required and may be seen as an
additional step in the above algorithm.
The deposit and data available
The copper oxide deposit considered in this study is located
in northern Brazil (Figure 1). It is approximately 2 km long
and 500 m wide and appears as a prominent NW–SE
aligned hill. The copper mineralization is hosted by copper
oxide minerals, mainly malachite, in a weathered
hydrothermal breccia and in a large weathering cap where
copper is disseminated and enriched in sheet-like mineral-
rich saprolitic units, as shown schematically in Figure 1
(bottom). Metallurgical studies indicate that iron and
potassium are key elements for predicting copper recovery,
indicating the need to evaluate all these three elements
throughout the deposit. The orebody mineralization model
374
supplied by project staff is a weathering envelope that hosts
oxide ore, Cu-oxide rich ore and Cu enriched in saprolite
ore, with saprolitic ore superimposed over mainly basic
volcanics. The deposit in this study is divided into two
units: Sector 11 and Sector 12. The available data include
654 RC and 310 DDH drillholes with samples analysed by
ICP-Plasma (Cu-K) and X-ray. There are 1136 five-metre
composites available in Sector 11 and 866 in Sector 12. The
descriptive summary statistics for Cu, Fe and K composites
for Sectors 11 and 12 are given in Table I. Figure 2 shows
the corresponding histograms for Sectors 11 and 12,
respectively.
Joint simulation of copper, iron and potassium
Normal-score transformation
Following the simulation steps using MAF described
earlier, a normal-score transformation is performed on the
Cu, Fe and K composites available in Sector 11 and Sector
12. Normal score transformations are based on rank
ordering of the data and decrease the influence of outliers.
This, in turn, assists the inference of the variogram and
estimation of covariance matrixes in the simulation process
that follows.
MAF transformation
The transformation matrix A MAF (Equation [1]) used to
generate the three min/max autocorrelation factors in
Sectors 11 and 12 is shown in Table II. MAF are calculated
by multiplying the vector of elements Cu, Fe and K by a
vector of loadings from the rows of the transformation
matrix. It should be noted that the MAF loadings are quite
different from the ones derived by PCA8. The lag ∆ in
Equation [4] used in this example is 20 metres and was
derived experimentally by testing several lag distances to
assure a suitable decorrelation and stable MAF
decomposition. Figure 3 shows examples of cross-
variograms between MAF from the present study that
demonstrate variable decorrelation. Experimental
variograms and cross-variograms for Cu, Fe and K are
shown and discussed in more detail in a subsequent section.
Variography of MAF
Variography on each MAF is performed. Figure 4 shows
the experimental and model variograms fitted to the three
MAF in Sector 11 and Sector 12. Note that all variogram
models are spherical. MAF variograms are subsequently
used in the simulation of each factor and the validation of
the MAF simulation results. MAF variograms show clear
spatial patterns, as expected. It should be noted that MAF
variograms are linear combinations of the variograms of the
original (normal score) variables.
Conditional simulation of MAF
Conditional simulation is performed independently on the
three MAF using a sequential algorithm 9 based on the
generalized sequential Gaussian simulation method10. The
simulations are performed on a grid of 320075 nodes within
the geological limits of Sector 11 and 239375 nodes within
Sector 12. Twenty simulations are generated in this study
for each of the two sectors and are validated in detail for
reproduction of data, histograms and variograms. The
validation of the MAF simulations is not presented here as a
subsequent section presents the validation of realizations in
the data space.
APCOM 2003
 Figure 1. Location map of the project and study area; and a schematic vertical section of the geological model of the deposit

Back transformations of MAF
The realizations of MAF were transformed back to
simulated normal score variables by multiplying a column
vector of simulated MAF in each grid node with the
corresponding inverse matrix of the MAF loadings in
Table II. Subsequently, the normal score Cu, Fe and K
realizations are back transformed to the data space, and
point support realizations are reblocked to 10 x 10 x 8 metre
blocks. Figure 5 shows three realizations of Cu through
horizontal sections and a vertical section of the deposit.

ASSESSING RISK IN GRADE-TONNAGE CURVES IN A COMPLEX COPPER DEPOSIT 375

 Table I
Descriptive statistics of 5-metre composites for Sectors 11 and 12

Cu % - Sector 11 Fe % - Sector 11 K ppm—Sector 11

Mean 0.41 Mean 10.93 Mean 14341.70
Median 0.24 Median 10.86 Median 11111.60
Std Deviation 0.59 Std Deviation 3.77 Std Deviation 14038.68
Kurtosis 75.83 Kurtosis -0.28 Kurtosis -0.52
Asymmetry 6.81 Asymmetry 0.29 Asymmetry 0.71
Minimum 0.01 Minimum 2.68 Minimum 123.00
Maximum 9.03 Maximum 25.04 Maximum 60300
Cu % - Sector 12 Fe % - Sector 12 K ppm—Sector 12
Mean 0.62 Mean 13.21 Mean 13715.20
Median 0.42 Median 13.70 Median 7679.90
Std Deviation 0.59 Std Deviation 3.65 Std Deviation 15334.67
Kurtosis 20.82 Kurtosis 0.11 Kurtosis 2.32
Asymmetry 2.98 Asymmetry -0.37 Asymmetry 1.44
Minimum 0.02 Minimum 0.01 Minimum 42.00
Maximum 7.46 Maximum 28.38 Maximum 93528

Histogram CU SECTOR 11 Histogram CU SECTOR 12

.500 Number of Data 811 Number of Data 1136
.250
mean 4096 mean .6183
std. dev. 5897 std. dev. .5892
.400 coef. of var. 1.4398 coef. of var. .9530
.200
maximum 9.0300 maximum 7.4550
upper quartile .4768 upper quartile .9045
median .2390 median .4250
.300
Frequency
Frequency

lower quartile .1280 .150 lower quartile .2000

minimum .0130 minimum .0180

.200 .100

.100 .050

.000 .000
.0 2.0 4.0 6.0 8.0 .02 1.02 2.02 3.02 4.02 5.02 6.02 7.02 8.02
CU CU

Histogram K SECTOR 11 Histogram K SECTOR 12

Number of Data 811 Number of Data 1136
.300
mean 14340.7000 mean 13715.2000
std. dev. 14030.0300 .300 std. dev. 15327.9200
coef. of var. 9783 coef. of var. 1.1176

maximum 60300.2000 maximum 93528.4000

upper quartile 25110.4500 upper quartile 22255.5700
.200 median 7679.9800
median 11111.6000
Frequency
Frequency

.200 lower quartile 1068.5000

lower quartile 848.4029
minimum 123.0000 minimum 42.0000

.100
.100

.000 .000
123. 10123. 20123. 30123. 40123. 50123. 60123. 42. 20042. 40042. 60042. 80042. 100042.
K K

Histogram FE SECTOR 11 Histogram FE SECTOR 12

.0700 Number of Data 811 Number of Data 1136
mean 10.9276 mean 13.2133
.0600 std. dev. 3.7717 .080 std. dev. 3.6454
coef. of var. 3451 coef. of var. .2759
maximum 25.0400 maximum 28.3750
.0500
upper quartile 13.3750 .060 upper quartile 15.7800
median 10.8800 median 13.7000
Frequency

Frequency

.0400 lower quartile 8.0505 lower quartile 10.9230

minimum 2.6800 minimum .0000
.0300 .040

.0200
.020
.0100

.0000 .000
2.7 7.7 12.7 17.7 22.7 0 10.0 20.0 30.0
FE FE

Figure 2. Data histograms of Cu, K and Fe in Sector 11 (left) Sector 12 (right)

376 APCOM 2003

 Figure 3. Cross variograms of MAF factors showing decorrelation

Table II
Transformation matrix AMAF for Sectors 11 and 12

Sector 11 -0.930 0.007 -0.126 Sector 12 0.704 -0.633 0.495

0.363 0.971 -0.400 0.697 0.773 0.008
0.057 0.238 0.908 -0.135 0.051 0.869

Factor 1- Sector 11 Factor 1 - Sector 12

1.50 1.50

1.20 1.20
Gamme (h)

Gamme (h)

0.90 0.90

0.60 0.60

0.30 0.30

0.00
70 140 210 280 350 0.00
80 160 240 320 400
Lag Distance (h)
Lag Distance (h)

Factor 2- Sector 11 Factor 2- Sector 12

1.50 1.50

1.20 1.20
Gamme (h)

Gamme (h)

0.90 0.90

0.60
0.60

0.30
0.30

0.00
120 240 360 480 600 0.00
90 180 270 360 450
Lag Distance (h) Lag Distance (h)

Factor 3- Sector 11 Factor 3- Sector 12

1.50 1.50

1.20 1.20
Gamme (h)

Gamme (h)

0.90 0.90

0.60 0.60

0.30 0.30

0.00 0.00
160 320 480 640 800 60 120 180 240 300
Lag Distance (h) Lag Distance (h)

Figure 4. Experimental and model variograms of the three MAF factors in Sector 11 (left column) and Sector 12 (right column)

ASSESSING RISK IN GRADE-TONNAGE CURVES IN A COMPLEX COPPER DEPOSIT 377

 Simulation 5
Cu%-Colour Profile
Simulation 10
Simulation 15
Figure 5. Selected conditional simulations of Cu for Sectors 11 and 12 for level #325 (left column) and cross-section 250NW (right column)
Validation of the joint Cu-Fe-K simulation results
Several validation checks are performed to assess the
results of the joint simulations of Cu, Fe and K using the
MAF transformations in both Sectors 11 and 12. Validation
involves calculation of histograms, experimental
variograms and cross-variograms of the simulated
realizations to ensure reproduction of original data
characteristics.
Figure 6 shows histograms from simulated realizations of
each variable in both sectors, and these can be compared to
the histograms of original data in Figure 2. Figure 7 shows
plots of variograms and cross-variograms for data and
simulations in Sector 11. All results suggest that the
reproduction of the original data spatial characteristics by
simulation is excellent. The reproduction of data
variograms and cross-variograms for Sector 12 is also
excellent. Recall that the variograms and cross-variograms
of original variables are not directly used in the joint
378
Simulation 5
Simulation 10
Simulation 15
simulation based on MAF, which used the variograms of
the MAF independent factors.
Risk assessment results and discussion
The mining project considered in this study represents a
US$120 million capital investment and aims to operate an
SX-EW copper recovery plant using traditional sulphuric
heap leach technology. One of the most important factors in
the economics of this project is the variability of
recoverable copper and the associated risk. Preliminary
metallurgical tests, summarized in Table III, indicate that
recoverable copper is controlled by the content of iron and
potassium, in addition to the original copper grade. This
highlights the importance of integrating the geological
model of the deposit with jointly simulated orebody
scenarios and metallurgical results so as to better identify
the possible variability of copper ore output (Cu tonnes)
from the SX-EW recovery plant.
APCOM 2003
 Histogram CU SIMULACAO SECTOR 11
.500 Number of Data 320075 Histogram CU SIMULACAO SECTOR 12
.250 Number of Data 239375
mean .4206
std. dev. .5515 mean .6004
.400 coef. of var. 1.3111 std. dev. .5050
.200 coef. of var. .8411
maximum 9.0300
upper quartile .5170 maximum 7.4600
.300 upper quartile .9079
Frequency

median .2400
.150

Frequency
lower quartile .1214 median .4460
minimum .0130 lower quartile .2028
.200 minimum .0180
.100

.100
.050

.000
.000
.0 2.0 4.0 6.0 8.0
.02 1.02 2.02 3.02 4.02 5.02 6.02 7.02 8.02
Back Transform
Back Transform

Histogram K SIMULACAO SECTOR 11 Histogram K SIMULACAO SECTOR 12

Number of Data 320075 Number of Data 239375
.300
.300 mean 13753.9100 mean 16340.8900
std. dev. 14070.9400 std. dev. 17805.6000
coef. of var. 1.0231 coef. of var. 1.0896
maximum 60300.2000 maximum 93528.4000
upper quartile 24640.5700 .200 upper quartile 25448.0400
Frequency

.200 median 9335.6410 median 11490.2300

Frequency
lower quartile 742.0432 lower quartile 1316.7340
minimum 123.0000 minimum 42.0000

.100 .100

.000 .000
123. 10123. 20123. 30123. 40123. 50123. 60123. 42. 20042. 40042. 60042. 80042. 100042.
Back Transform Back Transform

Histogram FE SIMULACAO SECTOR 12

Histogram FE SIMULACAO SECTOR 11 Number of Data 239875
Number of Data 320075 .0800
mean 12.7097
.0600 mean 10.3742 .0700 std. dev. 3.8296
std. dev. 3.6162 coef. of var. .3013
.0500 coef. of var. .3486 .0600 maximum 28.3750
maximum 25.0400 upper quartile 15.4126
upper quartile 12.7400 .0500
Frequency

.0400 median 13.2126

Frequency

median 10.1146 lower quartile 10.0738

lower quartile 7.4942 .0400 minimum .0000
.0300 minimum 2.6800
.0300

.0200 .0200

.0100
.0100
.0000
.0000 .0 10.0 20.0 30.0
2.7 7.7 12.7 17.7 22.7
Back Transform
Back Transform

Figure 6. Histograms for joint simulation of Cu, K, Fe using MAF. Sector 11 is shown on the left column and Sector 12 in the right

Given the above reasoning, the assessment of Cu average (e-type) from the simulations. It is apparent from
variability and risk in grade-tonnage curves for the deposit all grade-tonnage graphs that recoverable copper tonnage
is based on recoverable Cu. Recoverable Cu is calculated as variability is more clearly related to the in situ copper and
a function of the jointly simulated Cu, Fe and K contents potassium content than to iron content. Although there is no
for each block of the orebody models and the relationships restriction to a specific cutoff grade for Fe and K in the
derived from metallurgical tests, shown in Table IV. The metallurgical process, it is clear that for Fe-rich ores no
joint simulations described in the previous section provide wide variability in copper output is apparent in the grade-
the information needed to generate realizations of tonnage curves. This is supported by the geological model,
recoverable Cu at the support scale of 10 x 10 x 8 metre where Fe-rich grade zones are normally rich in malachite, a
blocks, and to subsequently generate grade-tonnage curves mineral with better kinetics in the metallurgical process,
for the deposit based on a range of Cu, Fe and K grades. compared with iron for acid consumption, and therefore not
The resulting grade-tonnage curves are then suitable for reducing copper recovery. Variability in recoverable copper
risk analysis as well as for ‘average type’ assessments for grade is, however, mainly dependent on the in situ copper
project economics. content. These results reflect the related metallurgical tests
Figures 8, 9 and 10 present the grade-tonnage curves for and the recovery functions developed, where the in situ
recoverable Cu as a function of Cu, Fe and K cutoffs. All copper content is the predominant control on recoverable
Figures plot the results from individual simulations and the copper.

ASSESSING RISK IN GRADE-TONNAGE CURVES IN A COMPLEX COPPER DEPOSIT 379

Figure 7. Reproduction of data variograms and cross-variograms of Cu, K, Fe, Cu-K, Fe-K and Cu-Fe in a simulation of Sector 1j note that
MAF uses only the variograms of the MAF independent factors

Table III
Preliminary metallurgical Cu recovery tests for the deposit under study

ORE TYPE NUMBER AVERAGE GRADE %

OF TESTS Cu (Head grade) Al K Mg Fe Fe2O3 Cu% (Rec) Rec%
SAPROLITE A 55 0.83 7.3 0.9 1.3 1.0 14.1 0.57 70.02
SAPROLITE B 14 0.28 7.8 1.2 1.0 0.6 16.8 0.09 35.49
SEMI-WEATHERED 9 1.2 5.1 0.4 2.3 4.0 6.43 1.01 84.65
SEMI-WEATHERED A 6 0.45 6.3 1.1 1.7 1.5 12.9 0.28 62.53

Table IV
Recoverable Cu equations from metallurgical tests and for different rock types

SAPROLITE A Cu Rec (%) = 0.05+0.93*Cu-0.018*Fe+0.007*K

SAPROLITE B Cu Rec (%) = 0.04+0.62*Cu-0.009*Fe+0.02*K
SEMI-WEATHERED Cu Rec (%) = 0.04+0.69*Cu-0.030*Fe-0.04*K

380 APCOM 2003

 400000.00 1.60

1.40

300000.00

Avg Cu (Recovered)
1.20
Tonnage Cu (t)-
1.00
Scaled

Grade %
200000.00 0.80

0.60

100000.00 0.40

0.20

0.00 0.00
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 1
Cutoff Cu%

Figure 8. Grade-tonnage curves for recoverable Cu as a function of Cu cutoffs. Thin lines represent individual simulations and thick dashed
lines represent the average (e-type) from the simulations

400000.00 0.70

0.60
300000.00

Avg Cu (Recovered)
0.50
Tonnage Cu (t)-

0.40
Scaled

Grade %
200000.00
0.30

100000.00 0.20

0.10

0.00 0.00
0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00
Cutoff Fe%

Figure 9. Recoverable copper tonnage as a function of iron content; results from individual simulations and average (e-type) from the
simulations

The grade-tonnage curves in Figures 8, 9 and 10, show
that the assessment of variability of recovered copper can
be investigated by integrating all three elements with the
geology of the deposit. This assessment of variability
provides a clear description of the inherent uncertainty of
potential copper outputs at the plant when analysing cash
flows and other financial aspects in the evaluation of the
present project. For example, consider a preliminary
operational Cu cutoff grade for the project of 0.6% Cu.
Figure 8 suggests that the output from the SX-EW recovery
process could range from 145000 to 210000 tonnes of
copper. With a planned annual Cu production of 45000
tonnes, the results suggest that the potential exists for the
project life to be substantially expanded.
At present, no restrictions to any specific cutoff grades
for Fe and K are considered in the evaluation of the deposit.
However, it is clear from Figure 9 that for the Fe-rich ores,
which are mainly composed of hematite and goethite
oxidized breccia zones, there does not appear to be wide
variability in recoverable copper outputs in the grade-
tonnage curves. This may result from the regular
occurrence of malachite with Fe-rich ores, which has better
kinetics in the metallurgical process, compared with iron,
for acid consumption.

ASSESSING RISK IN GRADE-TONNAGE CURVES IN A COMPLEX COPPER DEPOSIT 381

 400000.00
300000.00
Tonnage Cu (t)-
200000.00
100000.00
0.00
0 400 800
Figure 10. Recoverable copper tonnage as a function of potassium content; results from individual simulations and average (e-type) from
the simulations
Conclusions
Risk analysis in grade-tonnage curves for a complex multi-
element copper deposit in northern Brazil, or for any multi-
element deposit, being considered in a feasibility study is
critical for capital investment. This study presented the use
of an efficient joint simulation approach based on min/max
autocorrelation factors. The approach simplifies joint
simulations in mineral deposits by decorrelating variables
to independent factors that are then simulated
independently. The case study presented herein shows the
key practical aspects of the MAF approach and the
excellent validation of the results it generates.
The case study has presented three developments: (i) a
methodology for investigating uncertainty via jointly
simulating Cu, Fe and K; (ii) the construction of
recoverable Cu models as a function of solubility, given the
Fe and K content and the simulated realizations of the
deposit; and (iii) the quantification of related risk and range
of outcomes in Cu grade-tonnage curves. As a result,
valuable information is generated for the further analysis of
Cu grade and tonnage, solubility and recoverability effects
on the appraisal of the project.
Acknowledgements
This study had the support of Companhia Vale do Rio
Doce. Data of the project presented in this paper are
confidential. The results have therefore been selectively
scaled to avoid disclosure of information critical to the
project, but without affecting the approach developed in the
paper.
References
1. DIMITRAKOPOULOS, R. Conditional simulations
for the mining industry: Orebody uncertainty, risk
assessment and profitability in recoverable reserves,
ore selection, and mine planning. BRC, Brisbane,
2002. 355 pp.
382
0.50
Avg Cu (Recovered) Grade %
0.40
0.30
0.20
0.10
0.00
1200 1600 2000 2400 2800 3600 4000
Cutoff K (ppm)
2. CHILES, J.P. and DELFINER, P. Geostatistics:
Modeling spatial uncertainty. John Wiley and Sons,
New York, 1999. 695 pp.
3. GUTJAHR, A., BULLARD, B., and HATCH, S.
General joint conditional simulations using a fast
Fourier transform method. Mathematical Geology,
vol. 29, no. 3, 1997. pp. 361–389.
4. DAVID, M. Handbook of applied advanced
geostatistical ore reserve estimation. Elsevier,
Amsterdam, 1988. 216 pp.
5. DAVIS, J.C. Statistics and data analysis in geology.
John Wiley, New York, 1986. 646 pp.
6. WACKERNAGEL, H.J. Multivariate geostatistics.
Springer, Berlin, 1995. 256 pp.
7. GOOVAERTS, P. Spatial orthogonality of the
principal components computed from coregionalized
variables. Mathematical Geology, vol. 25, no. 3,
1993. pp. 281–302.
8. DESBARATS, A.J. and DIMITRAKOPOULOS, R.
Geostatisical simulation of regionalized pore-size
distributions using min/max autocorrelation factors.
Mathematical Geology, vol. 32, no. 8, 2000.
pp. 919–942.
9. DIMITRAKOPOULOS, R. Joint simulation using
min/max autocorrelation factors (MAF Factors),
Docegeo/CVRD internal report, 2002. 18 pp.
10. DIMITRAKOPOULOS, R. and LUO, X. Generalized
sequential Gaussian simulation on group size and
screen-effect approximations for large field
simulations. Mathematical Geology. 2003.
11. DEUTSCH, C.V. and JOURNEL, A.G. Gslib:
Geostatisitcal software library and user´s guide (2nd
Edition). Oxford University Press, New York, 1997.
369 pp.
APCOM 2003