See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/314297361

Drilling Grid Analysis for Defining Open-Pit and Underground Mineral

Resource Classification through Production Data

Chapter · March 2017

DOI: 10.1007/978-3-319-46819-8_18

CITATION READS

1 314

4 authors, including:

Roberto Menin
Glencore
6 PUBLICATIONS   2 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Roberto Menin on 20 June 2018.

The user has requested enhancement of the downloaded file.

Drilling Grid Analysis for Defining
Open-Pit and Underground Mineral
Resource Classification through
Production Data

Roberto Menin, Cassio Diedrich, Joao Dirk Reuwsaat,

and Wellington F. De Paula

Abstract The varied types of mineral deposits and geological features around the
world have led to the creation of a large number of techniques, methodologies, and
definitions for mineral resource classification. The most common methods used in
the mineral industry include kriging variance, drilling spacing, neighborhood
restriction, and conditional simulations. These methods generally do not use rec-
onciled production information, only long-term borehole information based on
personal judgment for defining confident intervals/limits on the mineral resource
classification. A drilling spacing back analysis study for defining mineral resource
classification was completed considering tonnages and grades confidence intervals
related to its respective production volumes, based on short-term production rec-
onciliation of analog deposits. The definition of adequate drill holes spacing and
detailed results for classifying mineral resources are demonstrated by both an open-
pit and an underground project adjacent to an existing mining operation. This study
has considered a Brazilian sulfide deposit (Cu-Au) operating mine as analog
information.

R. Menin (*) • J.D. Reuwsaat • W.F. De Paula

Vale Base Metals, 63 Grajau Street, 1st floor, Carajás, PA 68516-000, Brazil
e-mail: [email protected]; [email protected]; [email protected]
C. Diedrich
Vale Base Metals, 337 Power Street, Door #105, Copper Cliff, Sudbury,
ON P0M 1N0, Canada
e-mail: [email protected]

© Springer International Publishing AG 2017 271

J.J. Gómez-Hernández et al. (eds.), Geostatistics Valencia 2016, Quantitative
Geology and Geostatistics 19, DOI 10.1007/978-3-319-46819-8_18
272 R. Menin et al.

1 Introduction

New methodologies for classifying mineral resources are proposed every year with
constant efforts for reducing ore deposits estimation uncertainties using new tech-
nologies. However, given different types of mineralization and its geological
characteristics, it becomes impractical to define a single standardized methodology
for mineral resource classification. In general the estimation classification process
is based on qualified/competent person’s (QP/CP) judgment. Whether or not the
confidence of estimates (drilling spacing, experience, back analysis, or
geostatistical methods) are based on this judgment, their validity is subject to
assumptions that should be tested and validated against actual production results
through reconciliation studies (Morley 2003).
Several authors have reported the major issues associated with mining operating
cases that have not achieved planned production targets in its first years of operation
(Burmeister 1988). The main reasons are inherent to the mineral resources and
reserve classification processes, including errors associated with an inadequate
sampling procedure and lack of local geological knowledge (Harquail 1991).
Some studies present approaches for drill hole spacing using conditional simulation
(Dimitrakopoulos et al. 2009) or estimation variance calculation (Verly et al. 2014)
from long-term (hard data) drill holes information. Also a number of authors have
published a series of recommendations for quantitative estimates of accuracy to
classify resources and reserves (Vallée 1992; Wober 1993; Stephenson 1998). A
summary used for resource classification, by compiling public disclosure of mineral
projects issued by companies listed on the Toronto Stock Exchange, presents the
most commonly applied resource classification practices (Silva 2013). In regard to
these, none of the proposed mineral resource classification methods presents a
connection to actual operating aspects or makes use of an operational reconciliation
process. Therefore, mining configuration, geological knowledge, and production
aspects can help in the decision for better defining mineral resources categories.
In this paper an operating mining case study (sulfide Cu-Au deposit) has been
developed for defining mineral resources classification by using actual production
reconciliation results. The idea is to determine drilling grid spacing that best
support mineral resource categories for both open-pit and underground operating
processes. Different production volumes related to its respective geology and
operating practices were compared to actual exhaustive values for determining
confidence levels of mineral resources categories.
Production information of a copper mine operation, based on blasting reverse
circulation drilling, was used for both an underground and an open-pit projects due
to its similarity in geological aspects and production scale.
Drilling Grid Analysis for Defining Open-Pit and Underground Mineral. . . 273

2 Copper Mine Case Study and Methodology

A Brazilian large copper mine was used as analog information for both an under-
ground and open-pit sulfide Cu-Au project. This analog deposit will be called as
CMC (copper mine complex) and will be divided into two data sets, CM1 and CM2.
CM1 is the one defined for the underground project (UGP). The open-pit project,
referred to as OPP in the paper, took into account the CM2 data set.

The complete database of the CMC contains over 100,000 blast holes 1214 inch
analysis ranging from approximately 5 to 6 m average drilling spacing. A series of
different drill hole spacing grids were created for each mining method using
production information. A new geological interpretation wireframe was constructed
for each new grid by vertical extrusion of interpreted horizontal sections. Ordinary
kriging routines were used to estimate block models of each drilling grid for both
case studies.
The nearest neighbor interpolation algorithm was used for the construction of
each regular drilling grid. For the selection of the drilling grid spacing (e.g.,
30  30 m), a new block model was created with same expected drilling grid
dimensions. After running the nearest neighbor interpolation, only the closest
sample to the center of each block was selected and exported to create a new
regular drilling grid with approximate 30 m spacing based on original
production data.
Figure 1 illustrates how drilling grid selection was created and a sample selec-
tion was made for a regular sampling grid. The original geographic coordinates
(X, Y, and Z) of each sample was kept.
According to the geological interpretation for long-term operational assump-
tions, drilling grid supporting each production increment (global, annual, and
quarterly) was created according to long-term geological interpretation practices
and interpreted on both vertical and plan sections. Vertical section traverses were
created for supporting geological interpretation of plan sections without consider-
ing actual exhaustive information, only the information defined on the regular grids
by the sample selection procedure.
Fifty-three vertical sections striking N-S were interpreted with distance between
sections ranging from 10 to 40 m. A total of 21 plan sections, spaced at 16 m (bench
height) and between 128 and 192 m sea level, were interpreted in CM1. In the
same way, 21 vertical sections striking NE-SW were interpreted with distance
between sections ranging from 10 to 20 m. After modeling the vertical planes and
applying traverses for plan view interpretation, nine plan sections were interpreted
for CM2 considering 16 m bench height and elevation varying between 64 m and
192 msl (sea level). Figure 2 presents interpreted plan sections (1a and 2b) and
geological wireframes (1b and 2a) created for both CM1 and CM2 data sets.
The CMC operation long-term model (10  10  16 m block size) was used in
studied cases. Figure 3 presents the 20  20 m (grid spacing) interpolation results at
CM1 taking into account the geological interpreted mineralization. When different
statistical domains or geological interpretations are considered, the final copper
274 R. Menin et al.

Fig. 1 Sample selection procedure for a regular drilling grid

Fig. 2 Figures on top: (1) CM1 20  20 m interpreted plan section (left); CM1 10  10 m
geological wireframes (right); Figures on bottom: (2) CM2 30  30 m geological wireframes
(left); CM2 20  20 m interpreted plan section (right)

grades are defined by weighing the percent of each statistical domain within the
block.
A surface volumetric calculation for determining differences (tonnage, grades,
and metal) between drilling models in comparison with the production model was
Drilling Grid Analysis for Defining Open-Pit and Underground Mineral. . . 275

Fig. 3 Block model interpolation procedure

executed. Monthly and yearly topographic surfaces provided by operational teams

were utilized for all study periods.
Figure 4 presents a 4-year period represented by each annual topographic
surface.
An analysis of the OPP drilling grid spacing for mineral resource categorization
was made considering seven different sampling grids based on the CM2 production
data. The analysis of UGP drilling grid spacing for mineral resource categorization
was made considering six different sampling grids. The difference between inter-
polated grids and production models CM1 and CM2 was measured for each
specified period. Table 1 shows all sampling grids created for each of the case
studies.
At OPP, drilling grids spaces larger than 80  80 m were tested. However, these
provided poor results due to the nature (size and continuity) of mineralization and
were not included in this paper. At UGP, spaces larger than 100  100 m were not
tested due to lack of enough information at given drilling spacing.
CM1 data production was applied to the UGP mineral resource classification at a
high-grade cutoff value (0.75 % Cu) that supports the level of selectivity of an
underground mining operation.
As previously mentioned in the introduction, a number of authors have published
a series of recommendations for quantitative estimates of accuracy to classify
resources. Considering a systematic approach based on full operational knowledge
and geological features of both case studies, the following degree of confidence
limits will be applied to classifying mineral resources:
276 R. Menin et al.

Fig. 4 Annual comparison basis

Table 1 Drilling grids Drilling grids – CMC production data

CM1/UGP CM2/OPP
10  10 m 10  10 m
20  20 m 20  20 m
30  30 m 30  30 m
40  40 m 30  40 m
60  60 m 40  40 m
80  80 m 40  80 m
50  50 m

• Inferred mineral resource: a level of confidence of 15 % on the global recov-

erable metal content, tonnages, and grades
• Indicated mineral resource: a level of confidence of 15 % on the recoverable
metal content, tonnages, and grades over an area or volume corresponding to the
footprint of 1 year of production for a given deposit type in a mine or project
• Measured mineral resource: a level of confidence of 15 % on the recoverable
metal content, tonnages, and grades over an area or volume corresponding to the
footprint of one quarter of a year of production for a given deposit type in a mine
or project

3 Global Production Volume (Inferred Mineral Resources)

The level of confidence for the definition of inferred resources was applied in one
volume increment (global) for both CM2 and CM1. The tonnage produced in the
selected comparison period was obtained through volumetric calculation between
first and last year topographic surfaces. Figure 5 presents the percentage difference
Drilling Grid Analysis for Defining Open-Pit and Underground Mineral. . . 277

Global difference (%)

30.00%

15.00% Tonnes
ContMetal
relative difference

0.00% Cu%

-15.00%

-30.00%

-45.00%

-60.00%
10x10 20x20 30x30 40x40 60x60 80x80

Global difference (%)

30.00%
ContMetal
Cu%
Tonnes
15.00%
relative difference

0.00%

-15.00%

-30.00%
10x10 20x20 30x30 30x40 40x40 40x80 50x50

Fig. 5 (1) Global difference in terms of metal content (light blue), grade (blue), and tonnages
(red) for CM1 (top); (2) global difference in terms of metal content (light blue), grade (blue), and
tonnages (red) for CM2 (bottom)

between metal content, tonnages, and grades for all drilling grids in comparison
with the CM1 and CM2 production actual models.
The resulting tonnages, grades, and contained metal obtained in the estimated
models, for each drilling grid, had acceptable values for inferred resource classifi-
cation in comparison with the production exhaustive models. Tables 2 and 3 present
the percentage of scenarios within acceptable stipulated limits previously men-
tioned for both CM1 and CM2:
In the CM1 case, only the 80  80 m drilling spacing is not within acceptable
limits in terms of tonnages and metal content to the established production rate. In
the CM2 case, all drilling grids are within an acceptable range (percentage differ-
ences) for defining inferred mineral resources.
278 R. Menin et al.

Table 2 Global percentage Spacing (m) Metal (%) Tonnes (%) Grade (%)
of scenarios within acceptable
10  10 100 100 100
difference for CM1
20  20 100 100 100
30  30 100 100 100
40  40 100 100 100
60  60 100 100 100
80  80 0 0 0

Table 3 Global percentage Spacing (m) Metal (%) Tonnes (%) Grade (%)
of scenarios within acceptable
10  10 100 100 100
difference for CM2
20  20 100 100 100
30  30 100 100 100
30  40 100 100 100
40  40 100 100 100
40  80 100 100 100
50  50 100 100 100

4 Annual Production Volume (Indicated Mineral

Resources)

CM1 and CM2 global production were divided into five and four annual incre-
ments, respectively. This is in agreement with the annual production rates for both
open-pit (OPP) and underground (UGP) examples. The tonnage produced in the
selected comparison period was obtained through volumetric calculation between
the first and the last monthly topographic surfaces in the same year (January 1st–
December 31st).
The executed production rate on an annual basis ranged from 4.0–5.5 to
3.5–5.0 Mt for CM1 and CM2 open-pits, respectively. Figures 6, 7, and 8 present
a comparison of the percentage differences (grades, metal, and tonnages, respec-
tively) for each selected period for both CM1 (top) and CM2 (bottom), considering
the actual production information.
As commented, a level of confidence of 15 % on the annual recoverable
contained metal, tonnage, and grades is needed to support an indicated mineral
resource category. Tables 4 and 5 present the percentage of scenarios within the
acceptable stipulated limits:
The high-grade portions could not be mapped for drilling grid spacing higher
than 80  80 m in the CM1 case study. This reduced the tonnage of this material
significantly where only the grades were within acceptable threshold limits for
indicated mineral resources. Small differences in tonnages and grades were verified
in the 60  60 m drill spacing. The metal content was above the stipulated threshold
limits in two of the five annual increments. However, for these two scenarios, the
grade and tonnages were within acceptable limits.
Drilling Grid Analysis for Defining Open-Pit and Underground Mineral. . . 279

45.00% Annual Difference - Grade

30.00%
relative difference

15.00%

0.00%

-15.00%

-30.00%

-45.00%
10x10 20x20 30x30 40x40 60x60 80x80

30.00%
Annual Difference - Grade
relative difference

15.00%

0.00%

-15.00%

-30.00%
10x10 20x20 30x30 30x40 40x40 40x80 50x50

Fig. 6 (1) Annual difference in terms of grade for CM1 (top); (2) annual difference in terms of
grade for CM2 (bottom)

45.00%
Annual Difference - Contained Metal

30.00%
relative difference

15.00%

0.00%

-15.00%

-30.00%

-45.00%
10x10 20x20 30x30 40x40 60x60 80x80

Annual Difference - Contained Metal

30.00%
relative difference

15.00%

0.00%

-15.00%

-30.00%
10x10 20x20 30x30 30x40 40x40 40x80 50x50

Fig. 7 (1) Annual difference in terms of metal content for CM1 (top); (2) annual difference in
terms of metal content for CM2 (bottom)
280 R. Menin et al.

45.00%
Annual Difference - Tonnage

30.00%
relative difference

15.00%

0.00%

-15.00%

-30.00%

-45.00% 10x10 20x20 30x30 40x40 60x60 80x80

Annual Difference - Tonnage

30.00%
relative difference

15.00%

0.00%

-15.00%

-30.00%
10x10 20x20 30x30 30x40 40x40 40x80 50x50

Fig. 8 (1) Annual difference in terms of tonnages for CM1 (top); (2) annual difference in terms of
tonnages for CM2 (bottom)

Table 4 Annual percentage Spacing (m) Metal (%) Tonnes (%) Grade (%)
of scenarios within acceptable
10  10 100 100 100
difference for CM1
20  20 100 100 100
30  30 80 100 100
40  40 40 80 80
60  60 40 80 60
80  80 0 0 60

Table 5 Annual percentage Spacing (m) Metal (%) Tonnes (%) Grade (%)
of scenarios within acceptable
10  10 100 100 100
difference for CM2
20  20 100 100 100
30  30 100 100 100
30  40 100 50 100
40  40 75 75 75
40  80 25 75 50
50  50 50 100 50

In the CM2 case study, for the 40  40 m drilling spacing, slight differences in
the grade and tonnages than those stipulated were observed. However they reflected
a higher difference in the metal content compared to the values obtained by the
exhaustive production model. For two of the four annual production volumes, the
Drilling Grid Analysis for Defining Open-Pit and Underground Mineral. . . 281

30  40 m drilling grid scenario obtained tonnages which were slightly above the
stipulated limits; however, for these two scenarios, the grade and the contained
metal were within acceptable values.

5 Quarterly Production Volume (Measured Mineral

Resources)

CM1 and CM2 global production were divided into quarterly increments (20 and
15, respectively) which are in agreement with the quarterly production rate exe-
cuted. The tonnage produced in the selected comparison period was obtained
through volumetric calculation between the first and the last monthly topographic
surface of the quarter.
A level of confidence of 15 % on the quarterly recoverable tonnages, grades,
and metal content is required to support measured mineral resource categorization.
The quarterly production rates of CM1 case study ranged from 0.8 to 1.4 Mt. For
CM2 the ranges were from 0.75 to 1.5 Mt. The production period was divided into
actual monthly basis topography (short-term CMC) for the CM1 comparison.
Figures 9, 10, and 11 present a comparison of the percentage differences (grades,
metal, and tonnages, respectively) for each selected period for both CM1 (top) and
CM2 (bottom), considering the actual production information.
CM1 tonnages and metal content estimation results of 60  60 m and 80  80 m
drilling grid spacing are substantially different for the selected cutoff grade com-
pared to the exhaustive production model. It is more difficult to establish spatial
connectivity for the available samples when increasing drilling grid spacing, which
significantly reduces the volumes of certain portions of the deposit given the lack of
information for delineating mineralization. In the 30  30 m drilling grid spacing
scenarios, only a few quarterly production intervals lied outside acceptable metal
content limits in support of measured mineral resources. In the 40  40 m drilling
grid spacing scenario, there were several quarterly production intervals outside
acceptable tonnage and metal content threshold limits.
Minor differences in the tonnages and grades were observed in the CM2 case
study for a 30  30 m drilling grid spacing. These were verified mainly in scenarios
with lower amount of production information. The results demonstrated that dril-
ling grid spacing larger than 40  40 m does not predict within acceptable confi-
dence the production tonnages, grades, and metal content in a quarterly period.
Tables 6 and 7 summarize the percentage of scenarios within acceptable limits
defined for both cases studies.
282 R. Menin et al.

45.00% Quarterly difference on Grade

30.00%
relative difference

15.00%

0.00%

-15.00%

-30.00%

-45.00%
10x10 20x20 30x30 40x40 60x60 80x80

Quarterly difference on Grade

30.00%
relative difference

15.00%

0.00%

-15.00%

-30.00%
10x10 20x20 30x30 30x40 40x40 40x80 50x50

Fig. 9 (1) Quarterly difference in terms of grade for CM1 (top); (2) quarterly difference in terms
of grade for CM2 (bottom)

Quarterly difference on Contained Metal

30.00%
relative difference

15.00%

0.00%

-15.00%

-30.00%
10x10 20x20 30x30 30x40 40x40 40x80 50x50

45.00% Quarterly difference on Contained Metal

30.00%
relative difference

15.00%

0.00%

-15.00%

-30.00%

-45.00%
10x10 20x20 30x30 40x40 60x60 80x80

Fig. 10 (1) Quarterly difference in terms of metal content for CM1 (top); (2) quarterly difference
in terms of metal content for CM2 (bottom)
Drilling Grid Analysis for Defining Open-Pit and Underground Mineral. . . 283

Quarterly difference on Tonnes

30.00%
relative difference

15.00%

0.00%

-15.00%

-30.00%
10x10 20x20 30x30 30x40 40x40 40x80 50x50

45.00%
Quarterly difference on Tonnes

30.00%
relative difference

15.00%

0.00%

-15.00%

-30.00%

-45.00%
10x10 20x20 30x30 40x40 60x60 80x80

Fig. 11 (1) Quarterly difference in terms of tonnages for CM1 (top); (2) quarterly difference in
terms of tonnages for CM2 (bottom)

Table 6 Quarterly Spacing (m) Metal (%) Tonnes (%) Grade (%)
percentage of scenarios
10  10 100 100 100
within acceptable difference
for CM1 20  20 95 100 100
30  30 70 85 80
40  40 35 65 50
60  60 45 70 70
80  80 15 10 75

Table 7 Quarterly Spacing (m) Metal (%) Tonnes (%) Grade (%)
percentage of scenarios
10  10 93 100 93
within acceptable difference
for CM2 20  20 79 100 86
30  30 71 86 79
30  40 50 43 71
40  40 57 71 86
40  80 36 79 64
50  50 50 79 57
284 R. Menin et al.

Table 8 Final resource classification based on the back analysis study

Resource classification Measured Indicated Inferred
Underground project (CM1/UGP) 30 x 30 m 40 x 40 m 80 x 80 m
Open-pit project (CM2/OPP) 30 x 30 m 40 x 40 m 40 x 80 m

6 Conclusions

CM1 (UGP) study indicates that drill holes spacing wider than 60  60 m is
insufficient to adequately represent the mineralization and to reflect production
information due to lack of information at a given cutoff grade of 0.75 % Cu. The
historic reconciliation between the executed CMC production model and the CM2
(OPP) long-term model is not within acceptable threshold limits in areas with
drilling grid spacing wider than 40  40 m (measured + indicated).
Table 8 summarizes the drill hole spacing required to support mineral resource
categories for both CM1 and CM2 case studies using CMC production information.
The random selection of information can include waste information from neigh-
borhoods that are mostly mineralized. This randomness in the choice of information
can sometimes change the proportions of economic mineralization and waste in
some local portions of the model. This is part of the “real” variability of the deposit.
The mineralization characteristics of a deposit do not define the sample spacing
required for mineral resource categorization. A realistic production volume should
be considered in the evaluation of drill spacing for supporting mineral resource
categories. Realistic volumes are determined from analog deposits where produc-
tion information is available.
As mentioned in Diedrich et al. (2016), mining selectivity is understood as the
process of separating ore from waste, and its global and local concept is strongly
related to three mining functions such as geology, production rates, and mining
configuration that affect operational results. Operational aspects should be included
in a project or operation for better defining adequate drilling spacing and for
supporting mineral resources classification.

Bibliography

Burmeister B (1988) From resource to reality: a critical review of the achievements of new
Australian gold mining projects during 1983–1987. Macquarie University, Sidney
Deraisme J, Roth C (2000) The information effect and estimating recoverable reserves.
Geovariances
Diedrich C, Reuwsaat JD, Menin R, Paula WF (2016) Selection of operational mining unit (OMU)
size, Proceedings GEOSTATS 2016. Universitat Politècnica de València, Valencia, p 16
Dimitrakopoulos R, Godoy M, Chou C (2009) Resource/reserve classification with integrated
geometric and local grade variability measures. In: Proceedings orebody modelling and
 Drilling Grid Analysis for Defining Open-Pit and Underground Mineral. . . 285

strategic mine planning. The Australasian Institute of Mining and Metallurgy, Melbourne, pp
207–214
Harquail D (1991) Investing in junior mining companies. 6th Mineral Economics Symposium of
CIM. Canadian Institute of Mining, Mettalurgy and Petroleum, Montreal, p 10
Isaaks E (2004, 2005) The kriging oxymoron: a conditionally unbiased and accurate predictor. In:
Proceedings of geostatistics banff, 2nd edn. Springer, Banff, pp 363–374.
Jara RM, Couble A, Emery X, Varela EM, Ortiz JM (2006) Block size selection and its impact on
open-pit design and mine planning. J S Afr Inst Min Metall 106:205–211
Leuangthong O, Neufeld C, Deutsch CV (2003) Optimal selection of selective mining unit (SMU)
size. University of Alberta, Alberta
Morley C (2003). Beyond reconciliation: a proactive approach to using mining data. In: Large
open pit mining conference. Kalgoorlie, pp 185–192
Silva DA (2013) Mineral resource classification (NI 43–101): an overview and a new evaluation
technque. In: Centre for computational geostatistics. University of Alberta, Edmonton, p 307
Sinclair A, Blackwell G (2002) Applied mineral inventory estimation. Cambridge University
Press, Cambridge
Stephenson P (1998) The JORC code, 1987–1997, in Geology of Australian and Papua New
Guinean Mineral Deposits. In: Australasian institute of mining and metallurgy. Melbourne, pp
44–52
Vallée MA (1992) The guide to the evaluation of gold deposits: integrating deposit evaluation and
reserve inventory practices. CIM Bull 85:50–61
Verly G, Postolski T, Parker H (2014) Assessing uncertainty with drill hole spacing studies –
applications to mineral resources. In: Proceedings orebody modelling and strategic mine
planning symposium. The Australasian Institute of Mining and Metallurgy, Perth, p 10
Wober HA (1993) Classification of ore reserves based on geostatistical and economic parameters.
CIM Bull 86:73–76

View publication stats