Geoscience Reference
In-Depth Information
15
Conclusions
Abstract
Significant decisions are made on the basis of mineral resource estimates. There is sig-
nificant uncertainty associated with mineral resources because we sample a relatively
small amount of the deposit. The framework, techniques and numerical/statistical tools
that have evolved to address resource estimation in presence of sparse data and significant
uncertainty are summarized.
15.1
Building a Mineral Resource Model
prerequisites and getting on with the resource estimation to
meet the study objectives.
Most geostatistical studies are repeated as more data
become available or the objectives change. It is rare that a
particular geostatistical study is the first analysis of com-
pletely new data for a site that has never been modeled
before. It is important to assemble and review all relevant
prior work such as reports, maps, models, and data files.
Those that have studied the site in the past should be con-
tacted to avoid making preventable mistakes and to address
improvements that previous studies never had the time, data
or resources to address.
A generic workflow for geostatistics could be summa-
rized by eight steps. (1) Specify the goals of the study and
take inventory of the available measurements and concep-
tual data. (2) Divide the area/volume of interest into subsets
that are relevant for the specific situation. (3) Choose how
the mean of each variable depends on location within each
chosen subset. (4) Infer all required statistical parameters for
creating spatial models of each variable within each subset.
(5) Estimate the value of each variable at each unsampled
location. (6) Thoroughly validate the estimated model, en-
suring that the geologic and grade models are consistent with
the assumptions, data, domaining geology, and methodology
used in the estimation. (7) Simulate multiple realizations
to assess joint uncertainty at different scales. Finally, (8)
Post process the statistics, estimated models and simulated
realizations to provide decision support information. The
detailed implementation of these steps will depend on the
purpose of the study.
This section includes a summary of the steps involved in
building a block model, as discussed throughout the topic.
A typical work flow is summarized. The summary also
includes a review of the applications described, emphasiz-
ing the practical usefulness of the tools described, and the
(potential) benefits obtained by the owner/company.
More time is spent on getting ready to do resource model-
ing than on actually applying specific geostatistical tools. It
takes significant time to understand the geological setting,
the data, the study objectives and ensure that the modeling
workflow is designed to meet those objectives. Cleaning the
data takes a great deal of time. Often, the data are not
dirty
or incorrect, but the format is different and inconsistent, there
is missing data, there are different vintages of data, different
companies involved and so on. Preparing the site specific data
takes significant time. Understanding the geological context
of the data is essential to supplement sparse data and to make
good choices of model setup and modeling workflow.
Sufficient time must be allocated to sort out the study ob-
jectives, site specific data, analogue data and a conceptual
understanding of the site. Of course, time must be left to per-
form the geostatistical study and meet the study objectives.
Often, some data must be left out, some risk of error in the
database must be accepted and an incomplete understand-
ing of the geological context must also be accepted. Care-
ful documentation must be assembled of the data inventory
and the limitations that exist in the database and conceptual
understanding. There must be a balance between satisfying
