Mining Goes Digital – Mueller et al. (Eds)
© 2019 Taylor & Francis Group, London, ISBN 978-0-367-33604-2
184
Influence of drilling spacing on the mineral resources uncertainty
C.J.E. Silva, M.A.A. Bassani & J.F.C.L. Costa
PPGE3M-Post-Graduation Program in Mining, Metallurgical and Materials Engineering, Federal
University of Rio Grande do Sul, Porto Alegre, Rio Grande do Sul, Brazil
ABSTRACT: Rock cores from drilling works in mineral research activities are the main ore
modeling information and consume most of the budget. Data spacing should be optimized to
reduce uncertainties in the geologic model and mineral resource inventory. This paper investigates how the uncertainty of the mineral resources is influenced by distinct data spacings
through a case study on a bauxite mineralization in Brazilian Amazon plateau. The results
showed that the uncertainty of the mineral resource decreases as the data spacing decreases.
Depending on an acceptable uncertainty level, wider data spacings may be selected. This
optimal drilling data spacing may be defined using the uncertainty in the mineral resource
estimation, which is demonstrated in this case study.
1 INTRODUCTION
Mineral exploration aims at having a mineral resource statement, which plays a key role
to decide whether the mining project is feasible. These statements follow modern reporting
codes standards (CRIRSCO 2013). The Mineral Resources are classified as Measured, Indicated or Inferred depending on the confidence of their estimation. Furthermore, the codes
encourage the quantification of the risk and uncertainty of the mineral resources. Usually,
the uncertainty of the resource model is lower where the data are more densely sampled.
However, a denser sampling grid is more expensive. In this paper, mineral resource is defined
as the tonnage and grade above a given cut-off and we aim to quantify the relationship
between data spacing and the uncertainty of the mineral resources. This relationship helps
the project manager to select the most appropriate data spacing. The appropriate data spacing is that which results in an acceptable uncertainty of the resources and adheres to the
project’s budget.
2 BACKGROUND TO GEOSTATISTICS
Geostatistical simulation is a method for characterizing uncertainty in earth sciences modeling. It allows the generation of multiple equally probable realizations that honor the input
data, histogram, and variogram (Wild & Deutsch 2013). These realizations provide a measure of uncertainty about the phenomenon being modeled (Journel & Kyriakidis 2004).
The principle of sequential simulation, under a hypothesis of a multigaussian random
function, defines the Sequential Gaussian Simulation (SGS) (Isaaks 1990). When using SGS,
the data must be normally distributed. A normalization can be achieved by a normal score
transformation (Deutsch & Journel 1998). More details about mathematical proofs, parameters and applications of SGS can be found in several books (Goovaerts 1997, Deutsch &
Journel 1998, Journel & Ying 2001, Sinclair & Blackwell 2004).
Relevant works address the use of geostatistical simulation to measure the uncertainty and
its relation to data spacing (Wild & Deutsch 2013, Deutsch & Beardow 1999, Boucher et al.
2004, Journel & Kyriakids 2004, Pilger et al. 2001, Koppe et al. 2017). The methodology used
to develop our case study is similar to that used by Wild & Deutsch (2013).
© 2019 Taylor & Francis Group, London, ISBN 978-0-367-33604-2
184
Influence of drilling spacing on the mineral resources uncertainty
C.J.E. Silva, M.A.A. Bassani & J.F.C.L. Costa
PPGE3M-Post-Graduation Program in Mining, Metallurgical and Materials Engineering, Federal
University of Rio Grande do Sul, Porto Alegre, Rio Grande do Sul, Brazil
ABSTRACT: Rock cores from drilling works in mineral research activities are the main ore
modeling information and consume most of the budget. Data spacing should be optimized to
reduce uncertainties in the geologic model and mineral resource inventory. This paper investigates how the uncertainty of the mineral resources is influenced by distinct data spacings
through a case study on a bauxite mineralization in Brazilian Amazon plateau. The results
showed that the uncertainty of the mineral resource decreases as the data spacing decreases.
Depending on an acceptable uncertainty level, wider data spacings may be selected. This
optimal drilling data spacing may be defined using the uncertainty in the mineral resource
estimation, which is demonstrated in this case study.
1 INTRODUCTION
Mineral exploration aims at having a mineral resource statement, which plays a key role
to decide whether the mining project is feasible. These statements follow modern reporting
codes standards (CRIRSCO 2013). The Mineral Resources are classified as Measured, Indicated or Inferred depending on the confidence of their estimation. Furthermore, the codes
encourage the quantification of the risk and uncertainty of the mineral resources. Usually,
the uncertainty of the resource model is lower where the data are more densely sampled.
However, a denser sampling grid is more expensive. In this paper, mineral resource is defined
as the tonnage and grade above a given cut-off and we aim to quantify the relationship
between data spacing and the uncertainty of the mineral resources. This relationship helps
the project manager to select the most appropriate data spacing. The appropriate data spacing is that which results in an acceptable uncertainty of the resources and adheres to the
project’s budget.
2 BACKGROUND TO GEOSTATISTICS
Geostatistical simulation is a method for characterizing uncertainty in earth sciences modeling. It allows the generation of multiple equally probable realizations that honor the input
data, histogram, and variogram (Wild & Deutsch 2013). These realizations provide a measure of uncertainty about the phenomenon being modeled (Journel & Kyriakidis 2004).
The principle of sequential simulation, under a hypothesis of a multigaussian random
function, defines the Sequential Gaussian Simulation (SGS) (Isaaks 1990). When using SGS,
the data must be normally distributed. A normalization can be achieved by a normal score
transformation (Deutsch & Journel 1998). More details about mathematical proofs, parameters and applications of SGS can be found in several books (Goovaerts 1997, Deutsch &
Journel 1998, Journel & Ying 2001, Sinclair & Blackwell 2004).
Relevant works address the use of geostatistical simulation to measure the uncertainty and
its relation to data spacing (Wild & Deutsch 2013, Deutsch & Beardow 1999, Boucher et al.
2004, Journel & Kyriakids 2004, Pilger et al. 2001, Koppe et al. 2017). The methodology used
to develop our case study is similar to that used by Wild & Deutsch (2013).