Geoscience Reference
In-Depth Information
portant because a very small fraction of the mineral deposit
is sampled.
A second key concept is that the samples should be repre-
sentative of the volume (or material) being sampled, both in
a spatial sense and at the location where the sample is being
taken from.
Representative
means that the sampling and ana-
lyzing process used to obtain a sample results in a value that
is statistically similar to any other that we could have taken
from the same volume. Therefore, the sample values are con-
sidered to be a fair representation of the true value of the
sampled volume of rock. Representation in a spatial sense
implies that the samples have been taken in an approximately
regular or quasi-regular sampling grid, such that each sample
represents a similar volume or area within the orebody of in-
terest. This is often not the case and some correction will be
required. If the samples are not representative, then an error
will be introduced that will bias the final resource estimate.
In the context of data quality, the technical issues related
to sample collection can be divided into those related to field
work, and those related to processing of the information.
Some of the most important issues in the field include (1)
the location of drill holes, trenches, and pits; (2) the type of
drill holes used such as open-hole percussion, reverse cir-
culation, or diamond drill holes; (3) the drilling equipment
used; (4) the sampling conditions such as the presence of
highly fractured rock or groundwater; and (5) sample collec-
tion procedures. Core recovery or the sample weight should
be recorded. Geologic logging of the geologic characteristics
of the samples should be performed. Sample preparation and
assaying procedures are critical. The related quality assur-
ance and quality control program is a fundamental element
in the process.
Deposit- and mineral-specific sample preparation and as-
saying protocols must be derived and adhered to throughout
the sampling campaign. Heterogeneity tests (Pitard
1993
;
François-Bongarçon and Gy
2001
) are necessary to under-
stand sampling variances and minimize errors.
The construction and maintenance of the sampling data-
base requires a continuous quality control program, includ-
ing periodic manual and automatic checks. These checks
should be performed over all the variables in the database,
including grades, geologic codes, collar location and sur-
veys, and density data. Relational databases offer the pos-
sibility of easier data handling and improved quality control.
But they do not provide quality control by themselves, nor
do they replace the need for periodic manual audits.
deposit, the distribution of mineralized rock, and to develop
exploration criteria for increasing resources.
The level of detail in the geologic description of a deposit
steadily increases as the project advances through its differ-
ent stages. Economic factors are the most important ones af-
fecting the decision of whether or not to proceed with further
geologic investigations; therefore, most geologic work is ori-
entated towards finding more mineral resources, and to some
extent to more detailed general exploration.
Not all geologic information is relevant to resource esti-
mation. Geologic investigations for resource development
should concentrate on defining mineralization controls. Cer-
tain geologic details and descriptions are more useful for ex-
ploration in that they do not describe a specific mineralization
control, but rather provide guidelines for mineral occurrences.
The process of defining estimation domains amounts to
modeling the geological variables that represent mineraliza-
tion controls. The estimation domains are sometimes based
on combinations of two or more geologic variables, for
which a relationship with grade can be demonstrated. For
example, in the case of an epithermal gold deposit, an esti-
mation domain can be defined as a combination of structural,
oxidation, and alteration controls. In the case of a diamond-
iferous kimberlitic pipe, in addition to the geometry of the
pipe (lithology), internal waste relics are common, such as
granitic xenoliths. The frequency and volume of these within
the pipe may condition the definition of estimation domains.
The determination of the estimation domains to use is
based on geologic knowledge and should be supported by
extensive statistical analysis (exploratory data analysis, or
EDA), including variography. The procedure can take a sig-
nificant amount of time, particularly when all possible com-
binations of the available geologic variables are studied, but
it is typically worth the effort. Estimates are improved when
carefully constrained by geological variables.
The definition of estimation domains is referred to as
the definition of stationary zones within the deposit. An
important part of stationarity is a decision of how to pool
information within a specific zone within the deposit, within
certain boundaries, or the deposit as a whole. Decisions are
based on oxidation zones, lithologies, alterations, or structur-
al boundaries. The stationary domains cannot be too small;
otherwise, there are too few data for reliable statistical de-
scription and inference. The stationary domains cannot be
too big; otherwise, the data could likely be subset into more
geologically homogeneous subdivisions.
Defining the estimation domains in resource evalua-
tion is often equivalent to defining the mineralized tonnage
available in the deposit. Some units will be mostly mineral-
ized (with the potential of becoming ore), while others will
be mostly un-mineralized (almost certainly non-recoverable
low-grade resources or waste). The mixing of different types
of mineralization should be kept to a minimum to avoid
smearing grades across geologic boundaries.
1.3.2
Geologic Model and Definition
of Estimation Domains
Much geologic information is gathered during the investiga-
tions performed at different stages of a mining project. The
information is used to understand the genesis of the mineral
