Geoscience Reference
In-Depth Information
Adequate definition of the estimation domains is an im-
portant task for resource evaluation. Mixing of populations
within the deposit will generally produce a sub-standard re-
source estimate that underestimates or overestimates grades
and tonnages. It is very rare that any geostatistical technique
will compensate for a poor definition of stationarity. A good
definition of estimation domains means that only relevant
samples are used to estimate each location.
can be recovered and processed by mining. Any resource
evaluation, in order for it to become the basis for an eco-
nomic evaluation, has to be recoverable, and therefore in-
clude some dilution and ore loss. After applying constraints
derived from the ability to economically mine the deposit,
as well as all relevant types of dilution, the resource may
become a reserve.
Some resource estimators advocate the estimation of
purely geological in-situ resources, that is, an estimate of
the resources that are to be found if a snapshot of the de-
posit at the same scale and level of detail as provided by
the drill hole data and other geologic information could be
taken. Thus, it would be a description of its true geologic
nature, as it occurs at our scale of observation. This point of
view assigns to the mining engineer and economic evaluator
the task of converting the purely geologic resource into a
minable reserve. This is required to realistically describe the
economic potential of the deposit. In general, however, the
geologist and geostatistician (resource evaluators) are better
equipped to incorporate geologic dilution; otherwise, it may
go uncharacterized or poorly modeled.
Mining is a large scale industrial operation; selection of
large volumes is taking place over short times. Some mix-
ing of waste with ore and ore with waste is inevitable. The
failure to understand and properly estimate geologic dilu-
tion and lost ore explains most of the failures of resource
estimates. Although some degree of error or uncertainty is
expected, ignoring or mistreating knowledge of anticipated
dilution is an invitation for disaster. An interesting discus-
sion in layman terms about this issue can be found in Noble
(
1993
). In the context of using a block model to estimate
resources, the basic types of dilution often encountered can
be summarized as:
1. Internal dilution, related to the use of small size com-
posites to estimate large blocks, also called the volume-
variance effect. The more mixing of high and low grades
within the block, the more important this effect will be, as
is common for example with gold mineralization.
2. The geologic (or in-situ) contact dilution, related to the
mixtures of different estimation domains within blocks.
One reason for grade profile changes is the existence of
different geologic and mineralization domains. Mixing of
grades will occur when mining near to or at contacts.
3. The operational mining dilution that occurs at the time
of mining. The blasting of the rock is an important fac-
tor, since material shifts position. The loading operation
is also a source of dilution and ore loss since the loader is
never able to precisely dig to the exact ore limits.
An understanding of the information effect is also required.
The long-term block model is not used for final selection of
ore and waste. Rather, a different model is used to select ore
from waste that uses much more closely-spaced data avail-
able at the time of mining. In an open pit mine the mineral
1.3.3
Quantifying Spatial Variability
The grade values observed within a mineral deposit are not
independent from each other. Spatial dependency is a conse-
quence of the genesis of the deposit, that is, all of the geolog-
ical processes that contributed to its formation. The reader
is referred to Isaaks and Srivastava (
1989
) for an accessible
discussion on the subject, as well as David (
1977
), Journel
and Huijbregts (
1978
), and Goovaerts (
1997
) for more de-
tails.
A clear description of the spatial variability (or continu-
ity) of the variables being modeled is desirable. Knowledge
of the spatial correlation between different points in the de-
posit will lead to a better estimation of the mineral grade
at an unknown location. The spatial variability is modeled
using the variogram and related measures of spatial variabil-
ity/correlation.
A spatial variability model improves the estimation of
each point or block in the deposit. Parameters of the model
are important. Attention should be paid to the definition of
the nugget effect (the amount of randomness); the number of
structures; the behavior of the variogram model near the ori-
gin; and the specification of anisotropic features. Although
the spatial variability model will change depending on the
estimator and available data, it should be compatible with
accepted geologic knowledge. For example, the modeled an-
isotropies should be consistent with the spatial distribution
of known geologic controls, and the variances and ranges of
the models should be consistent with the overall variability
observed in the data.
Geologic variables have some degree of spatial correla-
tion. The challenges often encountered when quantifying the
spatial correlation lie with the inadequacy of the data being
used, inadequate definition of estimation domains, or use of
estimators that are less robust with respect to skewed data.
These challenges are discussed in detail in later chapters.
1.3.4
Geologic and Mining Dilution
In-situ
and
recoverable
resources must be differentiated. The
precise definition of recoverable varies in different parts of
the world. In general, the term refers to mineralization that
