Geoscience Reference
In-Depth Information
by obtaining the average TCu grade in 100 m slices defined
in each horizontal direction, and 20 m (two benches) in the
vertical direction. This was done both for the blocks and for
the declustered 10 m composites. Figure
14.29
shows the
drift of all the estimated block grades in the vertical direc-
tion, which compare well with the corresponding declus-
tered composites.
ing these checks, the full annotated sets of plan and sectional
views should be left as an historic backup, and for third par-
ties to analyze if needed.
This case study represents a very classical application
of the principles of mineral resource estimation docu-
mented in this topic. The following case studies illustrate
complementary and different techniques for special appli-
cations.
Volume-variance check:
The grade-tonnage (GT) curves
from the resource model were checked against the expected
GT curves, as predicted by the DG model, and correspond-
ing to the SMU-size blocks. Figure
14.30
shows the GT
curves for the resource model and for the DG-predicted
SMU model, for all the sulfide units combined (Domains
4-6). The correspondence is good, showing that, at a 0.5 %
TCu cutoff, the resource model has 5 % more tons and 9 %
less grade (is more diluted than predicted by the DG model).
A perfect match is not desirable, since the volume-variance
correction includes only internal dilution, and the expected
grade-tonnage curve is just another model. The resource
model should include more dilution than what is predicted
by the volume-variance correction, specifically geologic
contact dilution and some additional operational (planned
and unplanned) dilution.
14.2
Multiple Indicator Kriging, São Francisco
Gold Deposit
The São Francisco gold deposit is currently owned by Aura
Minerals, Inc., and is located in the State of Matto Grosso, ap-
proximately 560 km west of Cuiabá, in South-Central Brazil.
It is one of several gold deposits on an N-S trending belt.
The São Francisco (SF) deposit occurs in the Fortuna For-
mation of the Aquapei Group, which is composed of fine- to
coarse-grained meta-arenites, with meta-pelites, and occa-
sionally meta-conglomerates. The rocks in the area are fold-
ed, faulted, sheared and fractured within a series of broads
anticlines and synclines that can be traced over several ki-
lometers, with fold axes striking on a northwest-southeast
direction, and gently plunging north 5-10°.
Gold mineralization occurs in epigenetic quartz-filled
shear zones generally along the foliation that is oriented par-
allel or sub-parallel to the axes of the folds. It also occurs
in later, flat to shallow dipping quartz veins filling fracture
zones and cutting bedding and primary foliation of the host
rocks. Gold is frequently coarse, with nuggets several mil-
limeters in diameter in quartz veins. It also occurs within
limonite boxworks after pyrite and arsenopyrite. High-grade
gold mineralization also occurs where narrow 1-5 cm long
quartz veining is intense and it crosscuts bedding, producing
a type of stockwork mineralization.
Software check:
The software used for kriging, regardless
of whether it is a commercial package or an in-house devel-
oped software, should be checked in detail, to ensure that the
program actually does what it advertises. For example, an
alternative kriging program can be used to estimate several
blocks of the most important estimation domains, applying
the same correlogram models and kriging plans as used in
the resource model. Often this check is performed by third-
party reviewers and auditors, but it should be part of the vali-
dation protocol of every mining company.
14.1.29
Visual Validation of the Resource Model
14.2.1
Database and Geology
The resource model should be checked visually to ensure
that there are no inconsistencies in the estimation process
or other evident errors or omissions. Each estimated grade
in the blocks should be explained by composites surround-
ing the block, the correlogram models, and the kriging plan
used. When performing this check, it is advisable to keep in
mind the definition of the estimation domains, and whether
soft or hard boundaries where used.
To perform this check, several sets of plan views and cross
sections were plotted, showing the estimated TCu and SCu
grades for each block, the resource classification code, the
composites used in the estimation, and the three-dimensional
solids that represent the estimation domains. After complet-
As of December 2001, the database consisted of over 460
inclined and vertical surface drill holes. The sampling inter-
val is mostly 2 m down the hole, with some of the drill holes
inclined about 60° to the Northeast, while others are inclined
60° to the Southwest.
The geologic logging system implemented at São Fran-
cisco include descriptions for lithologies, sulfide content,
presence of quartz veins and sericitic bands, presence or ab-
sence of kaolin, hematite, gold nuggets, and the degree of si-
licification. The most significant characteristic is hydrother-
mal alteration, which is the basis for the geologic envelope
defined (HAZ envelope). Mineralization occurs when the
