Cutoff-grade analysis at Fazenda Brasileiro: mine planning for

declining gold prices

A. J. Wheeler and Ruy L. Rodrigues

Synopsis
Underground working at the Fazenda Brasileiro gold
mine started in 1988. Initially, surface declines served
both access and production needs, but as the mine
deepened the development of a central ore-hoisting
shaft became necessary. The effects of increased mining
costs in recent years have been made worse by falls in
the gold price. Steps that have been taken in response
to improve the long-term economic future of the mine
include the development of a declining cutoff-grade
strategy. The main objective of this strategy, for implementation over the life of the mine, is to increase its net
present value. Incorporation of an opportunity cost element in cutoff-grade calculations has led to higher
cutoff grades than were previously adopted. The application of a declining cutoff-grade methodology also
assists in short-term decision-making with respect to
major variations in the gold price. Additonally, operational cutoff grades have been calculated, for both
mine-limited and mill-limited situations, to aid the
evaluation of different situations. Finally, the overall
long-term planning model developed for the cutoffgrade study provides a means to analyse the sensitivity
of other mining parameters.
The Fazenda Brasileiro gold mine is in Teofilndia, Bahia,
Brazil, some 210 km northwest of the city of Salvador
(Fig. 1). Its development is summarized by the following
noteworthy dates:
1972
1974
1976
1984
1985
1988
1992
1999

Exploration for base metals started

Itapicuru Greenstone Belt outlined
Gold discovered
Open-pit and heap-leach plant start-up;
25 000 t/month oxide ore
Underground exploration started
Underground mine start-up (25 000 t/month);
carbon-in-pulp (CIP) process commissioned
Expansion to 80 000 t/month
Cumulative production reached 1 500 000 oz

Fig. 1

The build-up of gold production is charted in Fig. 2, which

clearly shows the increase in production during the past
decade.
Au, kg
6000

Manuscript first received by the Institution of Mining and

Metallurgy on 6 January, 2001; revised manuscript received on 21
February, 2002. Paper published in Trans. Instn Min. Metall. (Sect. A:
Min. technol.), 111, JanuaryApril 2002. The Institution of Mining
and Metallurgy 2002.

CIP + Heap-leach

5000
4000
Heap-leach only
3000
2000
1000
0

Fig. 2

The mine currently produces approximately 5000 kg gold

each year, which accounts for roughly 30% of the total
gold output of Companhia Vale do Rio Doce (CVRD). The
underground operations produce 1 000 000 t/year of sulphide
ore, which is fed to a CIP plant. There are approximately
1000 employees at the mine, including CVRD workers and
contractors. The mine has had a significant economic impact
on the region, which formerly depended on agriculture and
the production of sisal fibre.

Location map of Fazenda Brasileiro

Production of gold over mine life

Geology
The Itapicuri Greenstone Belt trends northsouth, extending
for 100 km with a width of approximately 40 km, in the
northwest of Bahia State, Brazil. This belt is composed of
three major lithological domains: mafic, felsicintermediate
and sedimentary. Granitic bodies intrude all of the sequence.
The Fazenda Brasileiro deposit is located in the southern part
of this structure, hosted within a sheared mafic unit of the
eastwest-trending Weber Belt. The deposit is mesothermal
and extends for some 13 km.
The Weber Belt is composed of a basal mafic unit and an
upper volcano-sedimentary unit. It is set locally between
basaltic rocks, with carbonatic sedimentary lenses of the
A35

Fig. 3

Metasediments

Volcaniclastics

Metagabbro

Chlorite schistsore zones

Pyroclastics

Gneiss

Metabasalts

Volcano-sedimentsCanto Sequence

Plan of southern part of Itapicuru Greenstone Belt

Riacho do Inco Sequence on the footwall and felsic volcanosedimentary rocks of the Canto Sequence in the hangingwall. The Weber Belt is also characterized by the presence of
intrusive rocks of gabbroic to felsic composition. These intrusions are associated with large structures where hydrothermal
alteration can be observed. Fig. 3 shows a plan of the south
part of the Itapicuru Greenstone Belt.
The main host rock for the mineralization is a magnetic
schist (quartzchloritemagnetite schist), which occurs in two
horizons to the south and north of the gabbroic unit. The
mineralization dips 4050 southwards. The mineralized ore
shoots also tend to have a plunge of approximately 10 eastwards. Six main ore shoots have been mined so far and
various minor satellite orebodies are known. These ore shoots
occur in a portion of the Weber Belt approximately 8 km in
strike length. The ore shoots are 200800 m long and 230 m
wide. At present, the underground workings are 500 m deep
and mineralization has been intercepted down to a depth of
800 m.
Initial exploration was generally based on surface diamond
drilling, employing a section spacing of 400 m along strike.
Definition of the most promising areas is accomplished by
systematic reduction in the interval between geological drill
sections from 400 to 100 m. In addition, the section spacing
of 100 m provides information with which to undertake the
design of primary access and infrastructure. More recently,
exploration drifts have been excavated to provide access to
the mineralization found at depth. Currently about 6000 m of
A36

waste development is completed each year.

As footwall development becomes available reserve definition is based on infill drilling, utilizing sections spaced at
25-m intervals. In some cases the spacing between sections
is reduced to 12.5 m. Diamond-drilling results are supplemented with scan-line maps of development drifts and
samples of development jumbo and production fan-drill cuttings. Access ramps are driven in the footwall to provide drill
bases for infill diamond drilling. Development of the stopes
and other mine infrastructure is undertaken from the access
ramps.
Geological data processing and mine planning make extensive use of the Vulcan software system. It is employed for
the interpretation of exploration information as well as the
design and subsequent processing of infill drilling data. These
data are then utilized for the design of ore development. After
excavation the roofs of development headings are mapped in
detail. Samples are also taken from the drill-holes of every
third development blast to confirm the grade of mineralization. After the completion of development work all available
information is used to reinterpret the geological model of
the mineralization. Whenever necessary fan drill-holes are
drilled at 25-m spacing and the final ore contour is defined
and the final block model generated. The solid model and the
block model are subsequently used by rock mechanics staff to
design the final stope, with the objective of minimizing dilution and maximizing recovery. Grades in the resource block
model are derived as interpolated drill-hole composite grades.

The grades originally come from fire-assaying of the drill-hole

samples. A kriging method of interpolation is employed.
Once the final stope design has been completed the design
of production drilling takes place, for subsequent execution
by the operations department. As stoping proceeds stope contours are surveyed using laser total stations to check the
actual recovery and dilution achieved. In this way it is possible to predict stope dilution and recovery factors and so
enhance future mining design and production plans.
Mine geology and mine planning processes are accomplished with the aid of the tools of the Vulcan software
system. Rock mechanics work utilizes the software packages
Examine2D and 3D and Phases, developed by the University
of Toronto, for numerical modelling of mine designs. In
addition, customized spreadsheets are used for specific rock
mechanics design functions.

trucks at loading points 100150 m apart. Modern techniques and state-of-the-art equipment are used, enabling high
productivity and standards of safety to be achieved. Getman
and Nitronobel auxiliary service vehicles are employed.
The sulphide ore is fed to a CIP plant, where it is crushed
in three stages and subsequently fed to one of two ball-mills.
The milled pulp is cycloned and fed to Knelson or Falcon
concentrators. The concentrate is cleaned on vibrating tables,
producing 50% of the final gold. Afterwards the concentrate
pulp is passed to a thickener, where process water is recovered. Thereafter, thickened pulp is transferred to two
cyanidation lines where dissolution of the gold takes place.
Gold is adsorbed on to activated carbon, which is screened in
elution columns, and the gold is released. The pregnant solution is pumped to an electrowinning section to recover gold
on cathodes. Gold bullion at 90% Au is subsequently produced in the form of 12-kg bars.

Underground mining operation

Cutoff-grade optimization
Access from the surface to the underground mine is provided
by ramps of 5 m 5 m cross-section with a gradient of 15%.
A vertical shaft was commissioned in early 1999 to facilitate
ore extraction from a depth of 400 m. Hoisting of run-ofmine ore is less costly than hauling it to surface via the access
ramp system. The shaft development is the first phase of the
long-term plan to extract minerals from depth and further
deepening of the shaft is planned.
The mining method is longitudinal sub-level retreat mining, employing remote scooptram loading. Sub-levels are
generally driven at vertical intervals from 13 to 15 m. Twoboom electric-hydraulic jumbos, typically HL205S Tamrock
rigs, are used for the development of stope access and sublevels. A schematic overview of the stoping operations is
shown in Fig. 4.

Fig. 4

Longitudinal sub-level retreat mining

The mine applies a practical minimum mining width of

3 m. The maximum stope width is approximately 30 m,
which relates to the true thickness of the orebody. If the ore is
greater than 10 m in horizontal thickness, separate parallel
panels are mined at about 12 m width. Atlas Copco Simba
H253 fan-drill rigs are used for production drilling. The
diameter of production drill-holes depends on the width of
the ore and the ring pattern. Drill patterns vary from 1.50 m
2.25 m for 64-mm holes to 1.75 2.60 m for 76-mm diameter holes. ANFO explosives are widely used. Emulsion
explosives are used for primers and wet holes. Non-electric
detonation caps are employed for the initiation of production
blasts.
Ore and waste loading is carried out by 6- or 9-yd3 Toro
400D and 501 loadhauldump vehicles. Hauling is performed by 20-, 30- or 40-t trucks (Volvo A-25 and A-35 and
Tamrock 40D). Ore and waste are mucked directly from
blast faces by loadhauldump vehicles and loaded into

The original mine layout was influenced by the surface outcrop of the deposit. Thus, development of the underground
mine started with a ramp system for access, services and
haulage. This configuration contained the capital expenditure
on underground access and infrastructure, thereby increasing
cash flows during the early years of production. However, as
the shallower orebodies were exhausted, deepening of the
mine workings led inevitably to extended haulage distances
and increased costs for development, extraction, ventilation
and general infrastructure. At the same time, exploration
from surface became increasingly difficult and more costly.
All these factors led to the development of the recently completed ore extraction shaft.
In addition to these challenges mining costs have been
continuously increasing and the gold price has been declining. These pressures have created the need to focus on key
aspects of mine planning. In particular, with the long strike
length of the deposit, the selection of ore blocks to be stoped
has a critical effect on the amount of developmentfor
required infill drilling access as well as for stope preparation
and subsequent mining. These problems have been further
compounded by the large proportion of inferred resources at
the mine and how these are best considered with respect to
long-term mine-planning decisions.
It was realized that a key element in addressing many of the
mine-planning issues was the cutoff grade. Accordingly, an
initial cutoff-grade study was completed at the mine towards
the end of 1996. This enabled an assessment of the potential
impact of cutoff grade and was followed up by further work
on cutoff grades, the results of which are described here.
Another initiative was started to gain a better understanding
of, and engineering control of, ore dilution. This, in turn, has
led to some of the results of the cutoff-grade study being
analysed with respect to variations in dilution.
The most significant aspects of this cutoff-grade project
were: (i) establishing maximization of the mine net present
value (NPV) as the principal objective in the calculation of
cutoff grades; (ii) use of an opportunity cost in cutoff-grade
calculation; (iii) consideration of a declining cutoff-grade
strategy as part of the life-of-mine plan; and (iv) ensuring that
cutoff grades were applied and tested against approximately
stope-size units of potential ore.
The way in which this cutoff-grade analyis was considered,
with respect to long-term mine planning, is presented
schematically in Fig. 5. The geological tonnage and grade
data were prepared so as to form an available reserve base.
A cutoff-grade analysis was then applied to determine which
parts of the deposit would be included in the long-term
A37

Fig. 5

Overview of cutoff-grade optimization

mining plan. This, in turn, could be used to assess primary

development requirements. On the basis of known mining
parameters a mine-life cash-flow analysis was then developed.
However, the cutoff-grade analysis itself is affected by the
predicted mine-life cash flow, as described in more detail
later. One approach to this chicken-and-egg situation, therefore, was to make initial assumptions and then to improve the
overall model through repetition of the different stages and
refinement of the various parameters. The objectives of this
project were to produce a life-of-mine plan and an accompanying, optimized cutoff-grade strategy.

Preparation of geological data

Blocking-out of resources
The general way in which resource data were handled in this
project is shown in Fig. 6. The initial part is the processing of
modelled geological resources. All the currently delineated
measured and indicated resources at the mine have been
modelled with the Vulcan software system. The mineralized
A38

structures were modelled initially within three-dimensional

wireframe envelopes. These wireframe structures stem from
the interpretation of potential ore structures on geological
sections, which were defined on the basis of drill-hole composites across structures having an average grade greater than
or equal to approximately 3 g/t. Block models were then
created within these wireframe envelopes. The individual
blocks measure 12.5 m (along strike) 1 m (across strike)
2 m (along dip). These blocks were also rotated so as to fit
in the best way with the overall 45 dip of the mineralized
structures.
For cutoff-grade purposes it is important that these defined
orebodies are broken down into approximately stope-size
units. Selection above or below a cutoff grade can then take
place at the same scale of selection as is used during the
mine-planning process. An efficient method was developed
using a customized Excel macro for the blocking-out of
resources on any supplied stope grid size. This was developed
using Excels Visual Basic for Applications (VBA). The way
in which orebodies were blocked out through this macro is

VULCAN

EXCEL

Fig. 6

Resource data handling

shown diagrammatically in Fig. 7. The first step was to cut

through the three-dimensional block model on the basis of a
stope block grid with supplied vertical and along-strike increments. Within each stope block the total volume, tonnage,
metal content and physical coordinate limits were calculated.
The macro also made checks for excessively small stope tonnages that may be created at the edges of each orebody. Such
small stopes were identified by their evaluated tonnage
being below a given threshold. It is incorrect to include such
stopes in the database for ore/waste selection on the basis of
their own contents, but as they can still represent an appreciable tonnage, it is incorrect to exclude them altogether. In
practice the sub-level stoping method is flexible enough to
have longer strike or vertical dimensions to deal with local ore
variations. A procedure was, therefore, developed to fold the
contents of these small stopes into the main orebody. When
these small stopes are encountered their contents are added
to the next adjacent full stope. This ensures that the complete orebody tonnage and metal content become available in
potential stopes for subsequent selection, with a selectivity
still based on stope-size units. In the cutoff-grade study a

stope grid size of 50 m (along strike) 20 m (vertically) was

applied, with a stope threshold tonnage of 5000 t.
The stope block evaluation macro ultimately determined
the following quantities for each stope block:
ID: identifier of each stope block, based on the name of the
parent orebody
EASTING: bottom left easting of stope block, on stope grid
system
ELEVATION: bottom left elevation of stope block, on stope
grid system
XMIN: minimum easting of model cells within stope block
XMAX: maximum easting of model cells within stope block
ZMIN: minimum elevation of model cells within stope block
ZMAX: maximum elevation of model cells within stope block
TONNES: total in-situ tonnes of each stope block
AU: average in-situ gold grade of each stope block
AREA: area of each stope block, as viewed from long section
in a vertical plane
THICK: average horizontal thickness of the material in each
stope block
A39

CLASS: number indicating resource classification of parent

orebody
The coordinate extents and the average stope block thickness were determined to assist in the subsequent calculations
of dilution and mining recovery. During this study separate
block models for 63 orebodies were output from Vulcan and
then blocked out by the process described previously. The

VULCAN

EXCEL

Table 1

Measured
Indicated
Total

Summary of stope block data

Number of orebodies
Mined Unmined Total

Number of stope blocks

Mined Unmined Total

12

101

12

32
19
51

44
19
63

101

104
172
276

205
172
377

number of stope blocks produced is summarized in Table 1.

Several of the orebodies supplied had already been mined
out, but it was considered important to input these data as
well for consideration of gradetonnage proportions in the
ore deposit as a whole and for extrapolation when considering
inferred resources. Another macro was set up to assign a
SECTOR number to different parts of the mine. This assignment could then be used at a later stage to apply different
cost and operational data to various regions.

Stope block data handling

Once a stope block resource table had been produced it was
important to be able to produce longitudinal sections (vertical
projections) of the mine that depicted these data, both for
application in this cutoff-grade study and for more general
use in long-term planning at the mine. An Excel macro was
developed to allow quick and efficient production of such
longitudinal sections. Input to this macro consists of the
stope block resource table, a colour map, easting limits for
the long section and elevation limits. All stopes were annotated with tonnage, Au grade and thickness and could be
colour-coded according to any of: AU, in-situ gold grade;
TONNES, in-situ tonnes; CLASS, resource classification;
THICK, horizontal thickness; or AU_CONTENT, contained Au metal in grams. Other data connected with
extracted stopes could also be used to shade stope blocks that
had already been mined out or those above or below some
applied criterion.
Various existing stopes were analysed with regard to the
planned and unplanned dilution ascertained from production
results. The findings were compared with respect to variations in thickness as well as the shape factor, used in the
Mathews method of stability assessment. For the purposes of
overall mine cutoff-grade analysis the following base case
parameters were determined: dilution = 33%; and mining
recovery = 89%. These figures enabled an in-situ cutoff grade
to be determined relative to the calculated at-plant cutoff
grades. Excel macros that provided evaluation against a single
supplied cutoff grade, evaluation against different cutoff
grades by sector, evaluation by stope list and the generation
of tonnagegrade curves and associated grade interval data
were developed to facilitate evaluation of the available stope
block resource data.

Cutoff-grade analysismethodology
Optimum cutoff grades

Fig. 7

A40

Stope block evaluation

For the purposes of cutoff-grade optimization it is important

to consider any mining operation as being composed of three
distinct phases, as shown in Table 2 for an underground
mining operation.1 This is because the most appropriate type,
and therefore calculation, of cutoff grade depends on which
particular stage is limiting the operation. Cost data should be
allocated according to the logical stage to which they belong.
The first stage is related to access; it depends mainly on the
physical size and location of the mineralized material at the
mine. At an underground mine the shafts and primary development come into this category. The principal ramp systems
are focused on access to certain areas of mineralized material.

Table 2 Main stages of underground mining operation, for

cutoff-grade analysis
Stage

Principal component

Operations

Access

Mineralized material

Shaft sinking,
primary development

Treatment
and stoping

Ore

Stoping, tramming,
hoisting
Crushing, grinding,
separating

Marketing

Metal

Smelting, refining,
selling

The amount (and therefore cost) of the primary development

is not directly affected by the amount of ore.
The second stage contains all activities associated primarily
with ore. This includes the mill treatment costs and for
underground mines also all stoping-related work, ore haulage
and hoisting. Most underground metal mines are usually
limited by their ore-handling capacity, i.e. their production
efforts are normally directed at keeping the mill at full capacity. The rate and amount of access work are varied as
required to ensure that the mill and ore-hoisting facilities are
kept at capacity.
The third stage contains those activities associated with
metal after it has left the milling processes. Gold mines are
almost never limited by this stage as there is usually no limit
to the amount of gold that can be sold.
Cost data were collated for individual activities at Fazenda
Brasileiro corresponding to these stages of the overall mining
operation. These data, prepared in Excel spreadsheet form,
were also linked with cash-flow models. This enabled the
parameters used in cutoff-grade determination to be directly
applied for the equivalent cash-flow analysis. For a milllimited mine the optimum cutoff grade is given by the
equation1
Gt =

V +

f +F
A

NSR r

V+
In situ Gtend =

f
A

NSR r

(1 + d)

(3)

Optimization methodology
Under normal conditions the opportunity cost is given by
F = c PV

(4)

where c is cost of capital (e.g. 12%) and PV is present value

of the mine.
This opportunity cost can be thought of as the interest that
could have been earned if the mines present value had been
deployed elsewhere. As the optimum cutoff grade at the present time is a function of the opportunity cost, which will
normally decrease from year to year, the optimum cutoff
grade will also decline with time. However, the initial opportunity cost will itself depend on the cutoff strategy adopted
through the life of the mine as the cutoffs affect both cash
flows and how many years of production are possible.
To determine an optimum cutoff-grade strategy, therefore,
an initial opportunity cost can first be estimated according to
an estimated mine life. This enables cutoffs and mine production data to be generated. From these the present value
and stope reserves for the mine can be estimated and then

(1)

where V is variable cost per tonne of ore for operations

directly associated with stoping, haulage and treatment; f is
fixed (time) costs per year, which are not directly affected by
the quantity of mineralized material, ore or metal produced;
F is opportunity cost, which is the change predicted over the
next year in the mines present value; A is mill capacity,
t/year; NSR is net gold price per gram (as the grade will be in
grams per tonne), after deduction of any required refining
and marketing costs; and r is recovery of the treatment plant.
This is the cutoff grade that should be used to decide whether
any potential stoping block should be brought into production. The Gt quantity described above is calculated in terms
of mined ore that will arrive at the mill. To calculate the
equivalent cutoff grade that pertains to in-situ ore the overall
mining dilution needs to be taken into account. Therefore
In-situ Gt = Gt (1 + d)

(2)

where d is overall planned and unplanned dilution, expressed

as a ratio between waste dilution and original ore material.
At the end of the mine life the present value of the mine
will be zero. Therefore, at that time the opportunity cost will
also be zero. The optimum cutoff grade at the end of the
mine life will therefore be

Fig. 8

Cutoff-grade analysis and underground design

A41

used in another iteration for the cutoff-grade calculations.2

This procedure is depicted in Fig. 8. Another important part
of this overall process is to ensure that the production data
correspond with the assumed mining and development parameters. For example, the production capacity must be
achievable with respect to the amount of primary development scheduled. During this cutoff-grade analysis these
different types of data were linked together by carefully structured spreadsheets so that changes could be made efficiently.
After a number of iterations the cutoff-grade strategy, reserve
base and corresponding economic parameters should all correspond. An optimum cutoff-grade strategy, therefore, will
normally have a declining cutoff grade from year to year.
Addition of the opportunity cost in cutoff-grade calculations
will normally suggest higher cutoff grades than would otherwise be the case. However, it will lead to the highest possible
mine present value. Another important advantage of using
the opportunity cost is in the handling of changes in cutoff
grade with price variations, as described below.

Initial stope selection

The optimum cutoff grade described above is used for initial
stope selection. The opportunity cost, F, used in its calculation should reflect current economic conditions, as already
described.

Gt1 =

V +

f+ F
A

NSR r

(1 + d )

Stope planning when secondary development already complete

Sometimes mining operations may have ceased after initial
(secondary) stope development has been completed. When
deciding whether or not such stopes should now be reopened
for production a different cutoff grade should be considered.
For this cutoff grade the component of the variable mining
costs that accounts for stope development should be
removed. This is because this money will already have been
spent, regardless of whether or not the stope actually goes
into production. Hence

Price change effects

The present value calculated for the mine will normally be a
function of the principal gold price being used for the majority of the mine life. If, in practice, the short-term gold price
suffers a major change that is considered temporary, it is possible to determine the most effective change to the current
cutoff grade. For example, consider the effect of the drop in
gold price in 1999, assuming it to be temporary:
Long-term gold price = $320/oz
Price in 1999 = $285/oz
Forecast to be produced in 1999 = 150 000 oz
Current PV assumed (based on $320/oz) = $70 000 000
Discount rate = 12%
Drop in revenue = 150 000 (320285) = $5 250 000
Current mine PV = 70 5.25/(1 + 0.12) = $65 310 000
If the present value of the mine is assumed to recover to the
same level a year later (with no appreciable drop in reserves),
the change in PV, dV/dT, is 70 65.31 = $4 690 000 and
opportunity cost, F, equals DV dV/dT where DV is interest
that could have been earned elsewhere and dV/dT is decline in
value due to deteriorating economic conditions. Hence
F = 0.12 65.31 4.69 = $3 150 000
Because of this lower opportunity cost the optimum cutoff
grade, as given by equation 1, will also be lower. This is confirmed by common sense. Periods with a lower gold price, a
lower cutoff grade and, consequently, a lower head grade will
result in less gold being sold. However, when the gold price
rises the cutoff grade should be raised (owing to the increase
in opportunity cost), resulting in more gold being sold at a
higher price.

(6)

In- situ Gt2 =

V2 +

f+ F
A

NSR r

(1 + d )

(7)

where V2 is variable ore costs, $/t, after stope development

costs have been deducted. This variable cost should then
include drilling, blasting, mucking, extraction and treatment.
In-stope caved muck
After or near the end of the extraction of planned ore from
any given stope additional caved muck may be present, which
could be left in place or extracted as ore. To make this decision the cutoff grade used should have had all the drilling,
blasting and other costs removed from the variable costs:

In- situ Gt3 =

V3 +

f+ F
A

NSR r

(1 + d )

(8)

where V3 is variable ore costs after all direct stoping costs

have been deducted. This variable cost should then include
mucking, extraction and treatment.
Development muck
Development muck is currently produced from a number of
different headings. Most of it is trucked to surface, regardless
of its contents. Therefore, to determine whether or not
such material should be treated as ore, only treatment costs
should be included in the variable costs for the cutoff-grade
calculation:
In- situ Gt4 =

V4 +

f+ F
A

NSR r

(1 + d )

(9)

Operational cutoff grades

The optimum cutoff grades described above relate most
directly to medium- and long-term mine planning. These
cutoff levels should be used to help to determine which possible stopes should be prepared for extraction. However, in the
day-to-day production environment, when various stopes are
already in production and muck is being produced in many
different development headings, there are also a number of
operational cutoff grades that can be applied in the pursuit of
maximum economic benefit (highest present value). Details
of the important operational cutoff grades identified in the
context of the different phases of mining operations at
Fazenda Brasileiro are given below.
A42

where V4 is variable ore costs associated with treatment only.

Mine-limited cutoff grades
All the operational cutoff grades described above are based
on the assumption that the mill is always operated at full
capacity, i.e. it is a mill-limited operation. From time to time,
owing to operational problems, the mill may be starved of
oreit will then be operating below capacity. This situation
is mine-limited and different (lower) optimum cutoff grades
should apply. When calculating mine-limited cutoff grades
fixed and opportunity cost components do not apply. Such
mine-limited cutoff grades will only apply to in-stope caved

Fig. 9

Tonnagegrade curves

muck and development muck, as it is hoped that long- and

medium-term mine planning are aimed at the mill being kept
at full capacity. The mine-limited optimum cutoff grade for
in-stope caved muck will be given by
In- situ Gb3 =

V3
(1 + d)
NSR r

V4
(1 + d)
NSR r

(11)

where V4 is variable ore costs associated with treatment only.

These different operational cutoff grades should therefore
be applied for the appropriate locations and operating situa-

Fig. 10

Cut-off grade analysisresults

Assessment of resources

(10)

where V3 is variable ore costs after all direct stoping costs

have been deducted. This variable cost should then include
mucking, extraction and treatment.
The mine-limited optimum cutoff grade for development
muck will be given by
In- situ Gb4 =

tions. Following this strategy will lead to the highest possible

present value of the mine.

Gradetonnage curves for the measured and indicated

resource stope blocks are shown in Fig. 9. The y axis of these
curves represents the total proportion of either the measured
or indicated tonnage above cutoff. There is clearly very little
measured material below a cutoff of 5 g/t. At cutoff grades
higher than 6 g/t the measured tonnage falls quite steeply.
The increase in grade with increasing cutoff is more gradual
for the measured stope blocks. For indicated resource stope
blocks there is a much sharper drop in available tonnage for
any cutoff grade between 1 and 10 g/t. For the same range in
cutoff grades there is a fairly uniform increase in the average
grade of stope blocks above cutoff. This information shows
that any imposed cutoff-grade strategy will have its greatest

Projected area versus cutoff grade

A43

effect on indicated stope blocks.

This is also supported by a combined graph of the total projected area (vertical plane) of available stope blocks above a
range of cutoff grades, as shown in Fig. 10. The total area of
indicated blocks available is smaller and declines over a much
wider range of cutoff grades. Fig. 11 shows the average (horizontal) thickness of available stope blocks against cutoff grade.
Fig. 12 shows the variation in total metal content (gold, g) with
cutoff grade. This shows, more clearly than any of the previous
plots, how much more gold is contained in the indicated stope
blocks and how changes in cutoff grade potentially have a
greater effect on the indicated blocks.

Optimum cutoff-grade strategy

For the base case set of resource blocks, using an overall
dilution of 33% and a mining recovery of 89%, a life-of-mine
reserve depletion and cash-flow analysis were calculated. In
this analysis the only imposed cutoff grade was 3 g/t on the
indicated resource blocks, consistent with the geological (and
previous mine-planning) cutoff used at the mine. This gave a
mine life of 10 years and a resultant NPV of $42 900 000.
For the same base case an optimum cutoff-grade strategy

A44

Fig. 11

Average thickness versus cutoff grade

Fig. 12

Metal content versus cutoff grade

was developed using the system described above for the

assessment of opportunity costs. This resulted in a cutoff
grade of 6.0 g/t being used for indicated stope blocks in year
2, followed by a cutoff grade of 4.8 g/t in year 3. For the
inferred stope blocks a declining cutoff-grade strategy ranging
from 5.8 to 4.8 g/t was developed. The result of this strategy
was an eight-year mine life, with approximately an extra
2 500 000 t of resource material being rejected. However, the
overall benefit of this approach was to increase the mine NPV
to $68 700 000an increase of 60%. The assumption was
made that such stope blocks at the higher earlier cutoff grades
can be made available in time, but even if practical difficulties
restrict the grade of stope blocks available, it is clear that
pursuing such a declining cutoff-grade strategy produces economic benefits.
For the purposes of sensitivity analysis these two operating
scenarios, with and without cutoff-grade optimization, were
repeated for a number of different modifications with respect
to the base case: Table 3 gives a summary of these results,
comparing cases with different amounts of dilution. Alongside
the calculated NPV the resultant average Au reserve grades are
given. The runs show that for a decrease, to 26%, in mining

Table 3

Operational cutoff grades

Effect of cutoff-grade strategy on NPV, $106

Non-optimized cutoff grade

Optimized cutoff-grade strategy

33% Dilution
NPV Au, g/t

26% Dilution
NPV Au, g/t

43
68

56
82

4.7
5.5

5.0
5.8

A set of operational cutoff grades, which applied to the current economic and operational conditions at the mine, was
calculated in accordance with the theory presented above
(Table 4). The opportunity cost used was that which
reflected the low gold price of $285/oz. When the current
Table 4

dilution higher present values are possible. A combination of

this dilution change and an optimal cutoff-grade strategy
almost doubles the original, unoptimized base-case NPV.
Similar alternative cash flows were produced by varying other
parameters, which included pessimistic what-if projection of
lower head grades; different dilutions for alternative mining
methods; and corresponding costs for higher-capacity bulkmining methods. The results from this work highlighted the
importance of focusing mining engineering efforts on the
reduction of dilution. These need to be supported by the
determination of any mining cost changes and a sound basis
for the subsequent prediction of dilution and mining recovery
for long-term planning purposes needs to be established.
The conclusions of this work to devise an optimized cutoffgrade strategy were as follows.
(1) Following a carefully planned declining cutoff-grade strategy will enhance the mines present value.
(2) A pessimistic scenario with respect to the mined grades
from the inferred resources deemed to be available would
render an optimized cutoff-grade strategy even more vital.

Operational cutoff grades

Type

Description

Initial stope
selection

Application of all
relevant stope
preparation costs
After stope
Stope development
development costs removed from
complete
variable costs/t
In-stope caved Drilling and blasting
muck
costs also removed
Development Only treatment costs
muck
in variable cost

In-situ cutoff grade, g/t

Mill-limited Mine-limited
4.7

n.a.

3.6

n.a.

3.4

1.3

1.2

0.9

gold price changes appreciably the operational cutoff grades

at the mine can and should be revised. These cutoff grades
are shown diagrammatically in Fig. 13. It is important that
the mine-limited cutoff grades are applied only in situations
when the mill is starved, i.e. operating at a level lower than its
capacity.

Surface

Shaft

In-stope caved muck

3.4 g/t
Variable costs,
transport and treatment

Development muck

After stope development

3.6 g/t
Variable costs, stoping,
transport and treatment

1.2 g/t
Variable costs,
treatment only

Initial stope selection

4.7 g/t
Variable costs, stope development,
stoping, transport and treatment

Fig. 13

Operational cutoff grades

(3) A decrease in dilution will also have an extremely positive

effect on the mines present value. Quantifying the potential
benefits should lead to a much better understanding of
the relationship between dilution control and mining costs
and stimulate an improvement in methods and engineering
control.
(4) Compared with the previous mine policy, use of an optimized cutoff-grade strategy will lead to a decrease in the
mined ore tonnage of approximately 30%, whereas the mined
gold will only decrease by approximately 14%.
(5) Decreasing the mining cost, at the expense of higher dilution, does not appear greatly to affect the mine NPV.
Testwork was also done to see whether different haulage
distances to different parts of the mine might require local
cutoff-grade variations. The calculated cost variations were
small and when combined with the other cost components in
the cutoff-grade calculations their effects were negligible.

Further work
An inferred resource was also used in the project, being based
on the assumptions of (a) a fairly uniform incidence of ore
occurrences with depth and (b) a distribution of minable
blocks similar to that of the existing measured and indicated
blocks.
It is known by geologists on site that initial exploration
drilling may well have missed occurrences of ore within the
CLX horizon. Mining and geological experience at the mine
supports the occurrence of many potential ore blocks at a
spatial frequency similar to that found in the upper regions of
the mine. To improve assessment of this inferred resource
base and its potential effects on the mines future the geological department intends to apply conditional simulation
techniques. Besides assisting in the development of cutoffgrade strategies it is intended that this will enable testing of
the primary development layout and schedule with reference
A45

to the locations of potential stope blocks. It is also aimed at

enabling an appraisal of the uncertainty associated with the
inferred resource estimation.
A detailed risk analysis of the parameters involved in this
cutoff-grade analysis and other aspects of the long-term mining plan has already been completed.
References
1. Lane K. F. The economic definition of ore (London: Mining
Journal Books, 1988).
2. Border S. N. Optimization of cut-off grades during design of
underground mines. Paper presented to Mining industry optimization conference, Sydney, June, 1991.
Authors
A. J. Wheeler Member graduated as a mining engineer from
Camborne School of Mines in 1981. He gained an M.Sc. in mining
at Queens University, Canada, and worked in Canada for five
years before returning to the United Kingdom to join Datamine
International. For the past eight years he has been an independent
mining consultant specializing in computer applications.
Address: Cambrose Farm, Redruth, Cornwall TR16 4HT, England;
e-mail [email protected]
R. L. Rodrigues graduated as a mining engineer from Ouro Prto
School of Mines in 1986. After initial experience with Companhia
Vale do Rio Doce (CVRD) and BP he worked as a mining contract
manager for five years before rejoining CVRD as chief engineer at
Fazenda Brasileiro, Brazil. E-mail [email protected]

A46