135
distribution as well as, on skew lognormal distributions, which are: 1) – porphyry coppergold deposit in Chile and Peru (Deraisme et al., 2008; Deraisme and Assibey-Bonsu, 2011;
Millad and Zammit, 2014), 2) – gold deposits, Australia (Assibey-Bonsu, 1998; Humphreys,
1998), 3) – iron ore deposit, Australia (De-Vitry et al., 2007) and 4) – Bauxita and Iron ore
deposits, Australia (Abazlov, 2014).
Another non-linear estimation technique, Indicator Kriging and different kinds of simulations as, Sequential Gaussian Simulation and Turning Bands Simulation can produce, in
general, a satisfactory recoverable resources model, however those approaches are still very
time consuming.
2 MINE GEOLOGY
Niobium mineralization within the Boa Vista project is associated exclusively with nelsonite
veins. These veins, comprising a magnetite-apatite-carbonate rock host the niobium mineralization in the form of pyrochlore. Nelsonite occurs as a complex stockwork of veins and
stringers with variable thickness and orientation, frequently in association with carbonatite
stringers and veins. The carbonatites are not well mineralized with niobium, but field relationships suggest that there are instances where nelsonite bodies appear to have fractionated
from original carbonatitic magmas.
The nelsonite stringers are hosted within a variety of pre-existing lithologies including
amphibolite and fenite, a metasomatically altered rock believed to originally comprise phyllite.
The geometry of the mineralized zone and the nature of the distribution of mineralized
veins result in a grade distribution in space that shows strong trends. Vertically within the
mineralized zone, the average Nb 2 O 5 grade increases with depth, as the trumpet narrows into
a more concentrated zone dominated by carbonatite and thick nelsonite dykes. The Nb 2 O 5
grade also decreases radially from the central feeder zone towards the limits of the drilled volume. These trends are also consistent with the results of exploratory de-clustering statistics;
the de-clustered average grade falls consistently within the increasing cell size, revealing that
the higher-grade central core zone has been more densely drilled than the lower-grade peripheral zones. The presence of these trends compel the use of a grade-related domain to delimit
a central mineralized zone where the niobium grades are sufficiently stationary to permit the
determination of reasonable structured interpretable variograms.
3 CASE STUDY
The objective of this case study is to access the mineable resources in a high complex Niobium
deposit, in a multivariate case using Uniform Conditioning (UC) and localize the panel result in
SMU scale using the Localized Uniform Conditioning (LUC). Additionally, the global results
and the annually mine plan results are compared against a Conditioning Simulation model that
enables to access the grade fluctuation in 100 equiprobable realizations. In the present case, the
Turning Bands (TB) algorithm is chosen for its accuracy and computational efficiency. As the
Nb 2 O 5 is the main variable and only one with economic interest, the Turning bands univariate
simulation is used to access the fluctuation grades and compare with LUC Nb 2 O 5 grades.
4 DESCRIPTION OF CASE STUDY DATA AND DE-CLUSTER
The area of the study is approximately 600 m (N-S) by 800 m (E-W) and the vertical extent
over which niobium mineralization has been identified is approximately 600 m. Two data set
are used in this study, the exploration diamond drilling (DDH) and the grade control reverse
circulation (RC) drilling. The exploration drillhole dataset is a majority drilling in spacing
between 25 × 25 m. and 50 × 50 m. The sampling intervals range from 1.25 m to 2.5 m. and
the grade control spacing is between 5 × 5 m. and 10 × 10 m. while the sampling in 1.25 meters
distribution as well as, on skew lognormal distributions, which are: 1) – porphyry coppergold deposit in Chile and Peru (Deraisme et al., 2008; Deraisme and Assibey-Bonsu, 2011;
Millad and Zammit, 2014), 2) – gold deposits, Australia (Assibey-Bonsu, 1998; Humphreys,
1998), 3) – iron ore deposit, Australia (De-Vitry et al., 2007) and 4) – Bauxita and Iron ore
deposits, Australia (Abazlov, 2014).
Another non-linear estimation technique, Indicator Kriging and different kinds of simulations as, Sequential Gaussian Simulation and Turning Bands Simulation can produce, in
general, a satisfactory recoverable resources model, however those approaches are still very
time consuming.
2 MINE GEOLOGY
Niobium mineralization within the Boa Vista project is associated exclusively with nelsonite
veins. These veins, comprising a magnetite-apatite-carbonate rock host the niobium mineralization in the form of pyrochlore. Nelsonite occurs as a complex stockwork of veins and
stringers with variable thickness and orientation, frequently in association with carbonatite
stringers and veins. The carbonatites are not well mineralized with niobium, but field relationships suggest that there are instances where nelsonite bodies appear to have fractionated
from original carbonatitic magmas.
The nelsonite stringers are hosted within a variety of pre-existing lithologies including
amphibolite and fenite, a metasomatically altered rock believed to originally comprise phyllite.
The geometry of the mineralized zone and the nature of the distribution of mineralized
veins result in a grade distribution in space that shows strong trends. Vertically within the
mineralized zone, the average Nb 2 O 5 grade increases with depth, as the trumpet narrows into
a more concentrated zone dominated by carbonatite and thick nelsonite dykes. The Nb 2 O 5
grade also decreases radially from the central feeder zone towards the limits of the drilled volume. These trends are also consistent with the results of exploratory de-clustering statistics;
the de-clustered average grade falls consistently within the increasing cell size, revealing that
the higher-grade central core zone has been more densely drilled than the lower-grade peripheral zones. The presence of these trends compel the use of a grade-related domain to delimit
a central mineralized zone where the niobium grades are sufficiently stationary to permit the
determination of reasonable structured interpretable variograms.
3 CASE STUDY
The objective of this case study is to access the mineable resources in a high complex Niobium
deposit, in a multivariate case using Uniform Conditioning (UC) and localize the panel result in
SMU scale using the Localized Uniform Conditioning (LUC). Additionally, the global results
and the annually mine plan results are compared against a Conditioning Simulation model that
enables to access the grade fluctuation in 100 equiprobable realizations. In the present case, the
Turning Bands (TB) algorithm is chosen for its accuracy and computational efficiency. As the
Nb 2 O 5 is the main variable and only one with economic interest, the Turning bands univariate
simulation is used to access the fluctuation grades and compare with LUC Nb 2 O 5 grades.
4 DESCRIPTION OF CASE STUDY DATA AND DE-CLUSTER
The area of the study is approximately 600 m (N-S) by 800 m (E-W) and the vertical extent
over which niobium mineralization has been identified is approximately 600 m. Two data set
are used in this study, the exploration diamond drilling (DDH) and the grade control reverse
circulation (RC) drilling. The exploration drillhole dataset is a majority drilling in spacing
between 25 × 25 m. and 50 × 50 m. The sampling intervals range from 1.25 m to 2.5 m. and
the grade control spacing is between 5 × 5 m. and 10 × 10 m. while the sampling in 1.25 meters