Grade control for Australian open pit gold mines

J. G. D A V I S
Department of Geology, University o f Wales College o f Cardiff, PO Box 914,
Cardiff CF1 3 YE, UK
Abstract: The deposits exploited in Australian open pit gold mines are often small and
of low grade with highly variable gold distributions. Consequently, they require detailed
grade control. The nature of these grade control programmes is outlined together with a
description of the various computational methods used. Particular attention is given to
geological input into the various stages of the grade control program so as to ensure the
reliability of sampling; to aid the production of bench extraction outlines; and to predict and
minimize dilution and ore loss. The application of geostatistical techniques in this respect is
also critically examined.

The past decade has seen the rise and in some specific case histories, the reader is referred to
cases fall, of Australian open pit gold mining. Collings et al. (1989), Richards (1990), Shaw
Throughout the 1980s numerous operations (1990a) and Devlin (1990), who have described
have exploited the upper weathered portions of procedures at Cosmo Howley, New Celebration,
deposits which had often previously been Mt Magnet, and Great Lady deposits respectworked by underground mining methods. This ively, and the proceedings Aus IMM Selective,
upsurge in gold mining activity took advantage Open Pit Gold Mining Seminar as well as other
of the carbon-in-pulp and heap leaching extrac- papers referred to in this paper. By way of an
tion technologies and was particularly signifi- overview, a more generalized approach is taken
cant in Western Australia.
here.
The following description is based mainly on
The nature of the grade distributions is often
erratic with a poor continuity of grade zones open pit mines operating in Western Australia,
along strike or down dip. The overall grade is though similar grade control techniques are used
also very low with many mills operating at feed elsewhere.
grades of around 3 grammes per tonne (g/t).
Bulk mining of the deposits would not provide
Geology of the deposits
a sufficiently high grade as large amounts of
The majority of gold production in Western
altered, but poorly mineralized, material would
inevitably be included with the ore sent to the Australia has come from epigenetic deposits
mill. Selective mining of small portions of mat- within the greenstone belts of the Yilgarn Block
erial using hydraulic excavators enables several with the Norseman-Wiluna Belt containing
stockpiles of different ore categories to be main- most of the larger deposits. Figure 1 shows the
tained, and then blended, to produce a relatively location of the state's geological provinces and
constant mill feed grade. The individual parcels principle gold deposits.
The gold deposits tend to be located within
of ore mined may be less than 50 banked cubic
metres (bcm) in size and so the feasibility stage the regional zones of brittle-ductile shearing and
drilling can not provide the level of detail the mineralization is also structurally controlled
required to identify the grade zone boundaries on a local scale and tends to be found in dilaon the pit floor. Consequently, the primary tional zones within subsidiary structures of the
occupation of pit geologists at these types of regional faults. The deposits are spatially associoperation is to oversee intensive bench sampling ated with intrusive granitoids, porphyries or
programmes and to interpret the results of these lamprophyres, but the mineralization tends to be
to produce bench extraction plans on a day to hosted by basalts, banded iron formations and
day basis.
Fe-rich sediments, all of which tend to have high
There are numerous descriptions of the grade 9 Fe/(Fe + Mg) ratios. The deposits are generally
control practices carried out at particular mines less than 1 km in length and less than 50 m wide.
in Australia and this paper attempts to present a
As well as being of amphibolite-greenstone
broad picture of the methods used along with an transition metamorphic grade the host rocks
appraisal of some of the computational aspects were silicified, carbonitized and suffered potassic
of the the grade control programmes. For alteration at the time of Au-Ag-As mineraliza-

From Annels, A. E. (ed.), 1992, Case Histories and Methods in Mineral Resource Evaluation
Geological Society Special Publication No. 63, pp. 219-232.

219

220

J.c.

DAVIS

Limits of
Yilgarn Block

Eastern

Western \
1 Gniess ~
Terrain

Greenstone belts

~k
,

Cross ~ ' /
~Province ~"
.J

100 Km
I

Wilun

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29

Boddington
Mr. Charlotte
Golden Mile
Edna May
Copperhead
Marvel Loch
Nevoria
Norseman
Mararoa-crown
Victory/Defiance
Kanowna
Porphry
Or(] Banda
Paddington
Coolgardie
Menzies
Sons of Gwalia
Harbour Lights
Sandstone
Lowlers
Sir samuel
Wiluna
Youanmi
Paddy's Rat
Reedys
Big Bell
Hill 50
Great F]ngall
Lancefield

Fig. I. Map of the Yilgarn Block showing the location of Western Australia's principal gold deposits.

tion. The gold is associated with iron sulphides

(pyrite, pyrrhotite, arsenopyrite) or occurs as
gold-silver tellurides or native gold.
Groves et al. (1988) have classified the structural styles of the primary gold deposits into the
following categories:
(i)
(ii)
(iii)
(iv)
(v)

shear-zone lodes;
telluride rich breccia lodes;
laminated quartz veins;
stockworks/quartz vein sets;
stratabound/stratiform replacement
deposits.

The shear-zone lodes are associated with

intense alteration and the mineralization is in
sheared rock while the native gold occurs in
quartz veins within shear zones. Examples
include the Golden Mile, which is the largest
deposit in Western Australia having produced
more than 1000 tonnes of gold, Ora Banda in the
Norseman-Wiluna Belt and Reedys in the
Murchison Province.
Many of the deposits mined underground
consist of discrete quartz veins which are con-

tinuous over short distances such as the Crown

lode at Kambalda. Stockwork ore, on the other
hand, is characterized by sets of thin discontinuous quartz veins hosted by highly altered
rocks. Mount Charlotte in Kalgoorlie and
Paddington are examples of this style of
mineralization.
Breccia lodes commonly contain gold telluride
mineralization and are typified by the very rich
Oroya shoot in the Golden Mile, Kalgoorlie.
On the other hand, BIF hosted deposits, such as
Hill 50 in the Murchison Province, fall into the
stratabound category.
Much of the gold production in the past ten
years has come from the weathered, upper
portions of these deposits. Figure 2 shows a sectional view of a typical weathering profile over a
primary gold deposit. The supergene gold tends
to have been dispersed into the wall rocks and
alternate zones of depletion and enrichment
give a pseudo-stratigraphic appearance to the
deposits. Native gold in these zones is associated
with goethite and hematite (the weathering
products of the iron sulphides). Remobilization
of gold in the supergene zones also leads to finer

OPEN PIT GRADE CONTROL

SAMPUNG
METHODS

MINING

WEATHERING
HORIZON

M~HODS

Earth Saw
Trenching

NATURE OF
MINERALIZATION

Surface
Laterite

Dozer Rip
Lines
Ditchwitch
Trenching

221

Remobilisation

Mottled Zone

-i.i.i'i.".

Leached
Soprolite

.........

Depleation

Enrichment

Saprolite
Water Table

211
,iiiii!iiiiiiiiiiiiiiiiii
:iqi:!%iiii!i!ii!!i!!!i!iiiiii!iiiii

Fresh Rock

Blast Holes

ID

._c

:i:!:ii:i:!i:!:!:i!:i:i:i:!:!:!:!!!!:i:i:i!:!!!:!:!:i:i:i:i:i:i:i:i:i:i:i:i:!:!:!:i:!:i:!:!:

iiiiiiiiiiiiiiiiiii!:i:iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii
Shear Zone
Lode

,--, I Secondary
. . . . Gold Holo

Fig. 2. Schematic section through the weathered zone of a gold deposit.

gold (lower silver content) and crystal growth

(Mann 1984).

Mining practices
With a few exceptions, most of the open pits in
Australia are less than 100m deep and l km
long. Typical mill capacities are around 1 million
tonnes per annum and the ore delivered to any
one mill may have come from a variety of
sources. A t Ora Banda in Western Australia, for
example, nearly 3 million tonnes of ore were
treated between 1985 and 1990. Most of this ore
came from the Gimlet South Pit but material
from 13 smaller satellite pits up to 55 km away
was also treated. Some of these deposits contained as little as 5000 ounces of gold.
Mining is usually carried out by contractors
using hydraulic excavators and dump trucks
with capacities of between 20 and 50 tonnes. The
pits are designed with bench heights between 10
and 20 m but each bench is mined in several thin
'lifts' known as flitches. These flitches are
between 1 and 2.5 m thick and allow more selective mining of the ore where strong vertical
variations in grade occur. This mining method is
applied in the upper portions of the deposits
where the down dip continuity of the lode can be

poor due to the development of the lateritic

weathering profile. Figure 2 shows the likely
variations in flitch height through the life of a
pit. Notice that the flitch height can be increased
in the sulphide zone of the ore body though it is
limited by the size of the equipment being used.
Table 1. Grade control ore categories for a Western
Australian deposit in 1990

Class

Grade range

Description

0.1-0.49 g/t
0.5-0.89 g/t
0.9-2.49 g/t
2.5-5.99 g/t
6.0-8.00* g/t

waste
mineralized waste
low grade ore
run of mine ore
high grade ore

0.5-0.99 g/t
1.0-1.29 g/t
>f 1.3 g/t

waste
sub-economic
sulphide ore

Oxide

Green
Blue
Yellow
Orange
Red

Sulphide
Green
Blue
Yellow

* Sample grades greater than 8.0 g/t are cut.

In order to maintain the mill feed grade, ore is
stockpiled by grade and by metallurgical type.

222

J . G . DAVIS

For example, Table 1 lists the grade ranges being

used for stockpiles at a Western Australian gold
mine in 1990. The the cut-off grades defining
each category are reviewed periodically to reflect
changes in both economic circumstances and the
relative proportions of each category of ore
encountered in the pit. The mineralized waste, or
sub-economic ore, category is stockpiled separately from the unmineralized waste as it may be
economic to process this material after the mine
production has ceased.

Grade control programmes

Grade control can be defined as the process
whereby the mill feed is optimized in terms of
maximizing gold recovery from the material
mined. This is achieved by minimizing dilution
and ore loss at the production stage of a mining
operation by the use of in-pit sampling to define
ore block boundaries.
Grade control involves the prediction of the
grade of ore stockpiles to allow efficient blending
and it provides information for reconciliation
between mill production and ore reserves estimates. The latter function enables remaining
recoverable ore reserve figures to be updated on
a regular basis and provides information on the
parameters to be used when evaluating new
deposits.
The highly variable nature of the gold deposits
described above means that they require closely
spaced sample data if the effects of dilution and

ore loss on the profitability of the operation are

to be reduced. These concepts are best explained
with reference to a diagram such as Fig. 3 taken
from Carras (1986). Dilution is partly caused by
the inclusion of waste material within the ore
body outline which is not detected by the sampling or is too narrow to be mined separately.
Ore loss also occurs in this way and additional
dilution and ore loss are the result of deviations
of the geologist's interpreted ore body outline
from its true shape. This 'edge effect' is unavoidable, and difficult to quantify, as it depends on
the degree of continuity of the ore body structure.
The stages which make up a typical grade
control programme are shown as a flow diagram
in Fig. 4. It is normally the task of mine
geologists to supervise these programmes and
their management has been discussed by Shaw
(1990b). The most important procedures involved
in the grade control methodology shown in Fig.
4 will be described chronologically.

Sampling
The effective implementation of selective mining
requires detailed information on the grade distribution of the bench/flitch which is to be mined.
This is provided by in pit grade control sampling
with, on average, one sample being taken for
every 30 tonnes of ore mined.
The collection of representative samples from
gold ore bodies is recognized as a major problem

Ore loss
Edge dilution
Internal dilution
Interpretted outline
True ore outline
Fig. 3. The concepts of dilution and ore loss (after Carras 1986).

Sample lines

OPEN PIT GRADE CONTROL

223

BENCHES CLEARED AND

MARKED FOR SAMPLING

TRENCHESDUG/I
HOLESDRILLED
I BENCH 4j.SAMPLINGF=:::I

'1

I SAMPLE ANALYSIS J

~,
I BENCH PLANS
PLOTTED

I CHECKSAMPLING

.,L--JIINTERPRETATIONOF I__~l ESlIMAllONOF [

ORE AND WASTE ~
INSITUGRADE
BOUNDARIES I I AND TONNAGE.,.
J-J.
BOUNDARIES
MARKED OUT
ON BENCH

I 1'

MINING BY
HYDRAULIC EXCAVATOR

POTENlqAL OR'E'"I
RESAMPLED l

IORD

TO STOCKPILES

Fig. 4. Flow diagram of a grade control program.

and stems from the complexity of their geology
and the variable nature of the gold mineralization. The fact that the concentration of the gold
within the host rock is very low yet samples vary
from less than one part per million to several
thousand parts per million, illustrates this point.
The presence of free gold particles in the ore
compounds these problems and creates severe
difficulties when assaying the samples due to the
physical characteristics of the metal.
Numerous articles have been written on this
subject and the reader is referred to Carras
(1987), for a discussion of the various components which contribute to errors in gold sampling in an Australian context. For the purpose
of this paper only the particular problems associated with each of the described sampling
procedures will be dealt with.
The main methods of providing sample
material in Australian open pit gold mines are
trenching (using dozer ripping or Ditch Witch
machines), blast hole sampling, and other
drilling methods (rotary air blast, reverse
circulation). Channel samples may also be taken
from active mining faces. Some, or all, of these
methods may be applied at any one mine and
Fig. 2 shows the vertical variation of the
methods within an idealized operation. Details
of the procedures involved and their relative
merits are discussed below.

Trenching
Most ore bodies lying above the transition zone
in the weathering profile can be mined without
the need for prior blasting. Sampling can therefore be carried out on material provided by
costeaning or trenching at regular intervals
across the strike of the orebody.
Initially, back-hoe excavators or dozers with a
claw attachment were used to score rip lines
across a bench thereby exposing in-situ material
from which channel samples could be taken.
This method has largely been superseded by the
advent of Ditch Witch machines.
The Ditch Witch machine consists of a large
chain-saw like cutter attached to the back of a
tractor fitted with low-profile tyres. It is capable
of cutting a trench about 15cm wide and
between 75cm and 150cm deep through soft
material and was originally designed to be used
for laying shallow cables and pipe lines. Archer
(1987) has described the introduction of the
machine to Australia from Oklahoma, and its
development for use as a grade control sampling
tool, during the mid-1980s. The first trials were
conducted at Marvel Loch Gold Mine in early
1985 and Ditch Witch trenching rapidly became
a widespread sampling method throughout
Western Australia and other states.
The reasons for the acceptance of Ditch Witch
machines at so many gold mines has been its cost

224

J . G . DAVIS

effectiveness compared to drill hole sampling

and its greater accuracy compared to rip lines
created with a bulldozer or back-hoe excavator.
The main advantages of Ditch Witch trenches to
dozer rip lines are as follows.
(i) The material produced is generally finer
grained and well mixed. The samples collected
can therefore be considered to be representative
of the material over the interval from which they
are collected.
(ii) As well as being relatively homogeneous,
the material is left on either side of the trench as
two windrows (see Fig. 5) whereas dozer rip lines
produce blocky trenches which are harder to
sample. The possibility of bias due to the selection by the samplers of the easiest material to
collect, is therefore eliminated.
(iii) The usual depth of penetration of the pit
floor with the ditch witch machine is around I m
compared to about 30 cm for a dozer rip line. If
the flitch height is limited to around twice the
trench depth then the ditch witch samples are
likely to give a better indication of the grade
over the entire flitch height.
(iv) The clean trenches produced make it
easier for the geologist to map pit benches/
flitches and to mark out ore zones prior to
mining.
The main constraint on the use of the Ditch
Witch trenching method is the hardness of the
rock to be sampled and of particular importance
are the presence of relic quartz veins and BIF
horizons in the weathered portions of the orebody. If hard material is intersected in the
trench, the boom on which the Alligator Chain is
mounted will lift up and thus avoid this material
therefore causing possible bias in the sampling
method.
The more robust Earth-Saw has therefore
tended to replace the Ditch Witch Chain at
many gold mines. It consists of a 2 m diameter
circular saw mounted on a tractor and is capable
of digging through harder material than a
machine of equal power but with a chain attachment. Other advantages of the Earth Saw are:
it produces a cleaner trench; the material displacement along the trench is much less (almost
none); there is less smearing of material along
the trench; the material is finer grained; the
trench produced has a constant depth and the
sample windrows are also more uniform in size.
The problem of sample displacement and
smearing of material along the trench is of
particular concern if there are sharp contacts
between ore and waste in the ore body being
sampled. The displacement can be seen when the

trench cuts across lithologies which have a

colour contrast and results from the digging
action of the chain throwing material forwards.
Material that is not deposited on the windrows
falls back into the trench and causes mixing of
the samples which results in the smearing of the
ore/waste boundaries. Both of these factors have
been reduced by customization of the design of
the Ditch Witch and Earth Saw machines for
their use as sampling tools.
These trenching methods provide a good indication of the positions of the large scale ore zone
boundaries in the cross strike direction and the
separation of the trenches is chosen in accordance with the variability of gold grades in the
strike direction.
The sampling procedure employed at various
mines using Ditch Witch Chain or Earth Saw
sampling has been described by Schwann (1987).
The following description is a schematic one
based on the general procedures used at several
operations visited in 1988 and 1990, in particular
the Ora Banda mine 65 km NW of Kalgoorlie
which is operated by BHP Gold Ltd.
Before the trenching is carried out the pit floor
is cleared of fly rock to prevent contamination of
the samples collected. This is carried out using a
grader. The positions of the trenches to be cut
are then surveyed and pegged out across the
bench at between 5 and 10 m intervals along the
strike of the orebody. Only the portion of the
bench which is expected to contain mineralization needs to be included and this area can be
identified using previous bench plans if the lode
zone cannot be visually defined.
When the Ditch Witch machine cuts the
trenches, it must always travel in the same direction across the flitch which, in most cases is in
the down-dip direction of orebody. This ensures
that the true positions of the samples can be
corrected for their displacement which is in the
order of 0.5 to 1 m.
Samples are collected at either 1, 1.5 or 2 m
intervals along the trenches. It is normal to
remove the top layer of the windrow before this
is carried out to reduce the contamination of
the samples. The samples are collected using
shovels, food scoops or split plastic tubes.
Roughly two kg of material is collected at each
sample interval and is bagged and then dispatched to the assay laboratories for analysis.
The locations at which the samples are
taken from the windrows varies from mine to
mine. Schwann (1987) states that the preferred
location, at mines exploiting stockwork style
orebodies, is at the outer side of the windrow
whereas lode style orebodies are sampled nearer
to the trench. These locations have been chosen

OPEN PIT GRADE CONTROL

225

DitchWitchMachine

AlligatorChain f

Smearing and
W"

utt

boundaries

i[ Trench.

Low sornple recovery

' ~~~,.i---

/ r

~,

~
A

I ~

~l..a

~
II t"

/
~

'~Pit Floor

9 S

]3

Ditch Witch

Sample
locations

Chain

iI
I

Earth

~.

Saw

15 cm

Fig. 5. (a)Profile of a Ditch Witch trench. (b) Cross section of a Ditch Witch and an Earth Saw trench (after
Schwann 1987).

226

J . G . DAVIS
material from each sampling interval from the
cyclone of the blast hole rig in a large plastic
bucket and then reduce this down to the desired
sample size using a riffle splitter or spear
sampling device. Some mines use autosamplers
attached beneath the cyclone which collect a set
fraction of the material from each interval
drilled, however, these devices can be prone to
bias due to sizing the sample material as it is
collected. Also, there is a potential for downgrading of samples if excessive amounts of drilling fines are lost from the sample collection
device. This problem can be significant when
gold is associated with brittle minerals, such as
pyrite, which are finely ground by the pulverizing action of the blast hole drill.
The usual problems associated with open hole
drilling also occur with blast hole sampling.
These include smearing of grade down the hole,
poor sample return in fractured ground, the
accumulation of denser material at the bottom
of the hole and variations in hole width where
the competency of the ground changes.

on the basis of repeatability amongst groups of

samples from the trenches and, to some extent,
on reconciliation of the sample grades with the
mill feed grades. The risk of bias in the sampling,
due to size fraction of the crushed material in the
windrows, is therefore recognized.

Blast hole sampling

In the drill and blast situation, sampling the
cuttings from the blast holes provides the most
cost effective means of grade control. Hydraulic
or compressed air powered blast hole rigs are
used to drill a tightly spaced array of vertical
holes. When mining ore, these holes are typically
6 m in depth allowing two 2.5 m flitches to be
broken together with a 1 m sub-drill. The average spacing of the blast hole grid is around 5 m
by 5 m but a table listing the spacing used at
several operations in Western Australia is given
by Bettenay & Shaw (1990).
In almost all cases, samples are collected from
the drill rig as the hole is advanced rather than
later sampling the spoil heaps around the hole
collars. One sample may be taken representing
each flitch or at regular one metre intervals
down the hole. As the hole diameter will be
between 79mm and 150mm, the mass of
material to be sampled from each hole can be
around 20kg and therefore needs splitting
before a 'representative' sample of between 2
and 5 kg is collected. It is usual to collect all the

Grade control drill&g

The ability of blast hole drilling to provide
samples from several successive flitches has
led some operations to use this method to
supplement or replace trenching in the free-dig
situation, If this is the case, then the holes may
be inclined allowing the geologist to more accu-

BENCH PLAN 7 6 0 RL
9-,.~'1~1
,

I
I
-.
--=/~ "~-~'~P~'I I
I
I
I
I
i
I I I ~ "

-~;~
~

'I'I'

I
111

l~l"--r
i
I~I
I
I
I
I
I~I
I
II
I _ II'MI
I
I
I
i
I ~J
I
I I
I
I
I~i
I
I~l
I
I
I
I
I
,
I
,
I
I
I
I

'

~IN"=i~t I I
I"~'N,-=,.I
I I I i'~IPl
I
I ~ I
I
l
I
I
I
I
I
I
I ~
I
I
I
I
I
1 I
I
I
I
I _
i
I
I ~
i.l
I
I
I
I
I I I
I
I
I~-~
ll~'~i
III'II
I
I
I
I
I~I
--~ i
I ~.~
I
I
I
I
"~g
I
I
1
I
I
I
-" ~
. ~
~
'
%
1
I
I
~+
I

~
I ir~

; , ' ~ - ~ ,

+
+

4"

4-

~[Skorn

R
[

Cronod,oriteI/I

"~

4,

..........

Andesite
limestone

4.

i|

Ore

+
+

.~

,*+%- .+ . . + . + . + .

.~

4"

"F

l~ILS,~
i i ~
I~!1
~
I
I!~1
lml
1431
I~l I

i
~ .
I ~1~o
i,e~i
i
,
r~l
I
I
[
I
I
I
I
I i /
I
I
I

~-04"

4,

l
I
I
I
I
I
I
I
I
IN,e-=~l
I
I
I
I
I ~'~=-mlL=~'~ I
|
i
I
~
I
I
I
~
I
I
i
i,.-.J",~--I
I
I
I
I ~'--4 I ~,![
....
--~'_--I
I ~
T'~S~
.,..,.

"~

.~
t

"

~F

"::-"

.~

"'+:+ " *~+++++:++:++-'+'+'+'+'+'+'+'~

Cloy Brec ... . . . .. . . . . . . . m m. - I m.- l m. . . I . . . . . %:* ++*:*~I~l

Metres

F~176 1
:':~ :41

Fig. 6. Geological bench plan from Browns Creek, New South Wales.

4.

OPEN PIT GRADE CONTROL

rately determine the horizontal boundaries of
the ore zones in each flitch.
Some pits use vacuum and reverse circulation
(RC) drilling for all grade control sampling.
Shaw (1990a) details the use of RC drilling
at several of Metana Minerals open pits in
the Murchison Province of Western Australia.
Reverse circulation is the preferred sampling
method as the problems of sample loss and
contamination are reduced compared to vacuum
or rotary air blast (RAB) drilling.

Truck dump sampling

At some mines, individual truck lodes are
dumped separately and re-sampled before they
are reloaded and delivered to ore or waste stockpiles. This is carried out when the geologist is
uncertain of the grade of a particular area of the
bench due to spurious results from the in-pit
sampling or if material, which looks promising,
is discovered while waste is being mined.
Each truck dump of between 30 and 50 tonnes
is dispatched to a stockpile or waste dump on
the basis of the results of one grab sample
collected from it. The sample tends to consist of
the more friable material as this is easier to
collect but a biased result is likely if there is a
contrast in the strengths of mineralized and
unmineralized material. Duplicate sampling of
truck dumps often indicates a poor repeatability
of assay grades and so truck dump sampling is
avoided if possible.

227

Bench mapping
The level of in-pit mapping carried out varies
from mine to mine. In general, pit geologists aim
to produce a bench geological plans at l : 250
scale and at 5 m vertical intervals, but this is not
always achieved due to production pressures.
The benefits of geological mapping in open pits
have been explained by Watchorn (1990) and he
states that routine, detailed geological mapping
is only carried out at 10% of the open pits in
Western Australia. By contrast Fig. 6 shows a
bench geological plan from a skarn-hosted gold
deposit in New South Wales.
When producing dig plans for the excavation
of ore zones in the pit, the geological trends of
the lode structures become as important as the
grade control sampling and these can only be
re~,ealed by detailed geological mapping. The
level of mapping carried out at most mines is,
however, only aimed at identifying the larger
scale gold-bearing structures.

Bench grade control plans

The aim of the in-pit sampling and mapping is to

aid the production of bench grade control plans
to which the outlines of ore and waste blocks can
be added. These outlines can then be used to
mark out the bench prior to excavating the ore.
The grade control plans show blast holes or
trenches which are colour-coded, on the basis
of the assay results, by manual or computer
methods. If geological logging of the samples is
undertaken then the lithological codes assigned
to each sample may also be included. The pit
Assaying
geologist groups together areas of the bench
The more remote gold mines tend to carry out which are felt to be of a consistent grade within
sample analysis on-site but most mines use one of the ore stockpile categories. If several
commercial laboratories which handle many metallurgical types of ore are present in the same
thousands of samples per day. Typically the bench, such as oxide and sulphide ore, then the
samples are crushed, ground and split to provide boundaries between these will also have to be
a 50 gramme charge for fire assaying. Replicate included on the plans. The minimum size of the
samples can be used to monitor the performance shapes produced determines the selectivity of
of the assay laboratories but bias in the sample the mining method and the confidence in the
preparation procedure can only be detected individual assay results.
Figure 7 shows a bench grade control plan
using standards of a known grade. The relative
merits and weaknesses of various assaying and from an open pit gold mine in Western Austrasample preparation techniques with respect to lia. The geologist's interpreted ore outlines and
gold samples is a topic which has been discussed the boundaries between different ore types are
by many authors including Royle (this volume) shown with assay results from samples taken
and Sawyer (this volume). A lower standard of from Ditch Witch trenches.
An alternative to the two-dimensional apaccuracy is generally accepted for grade control
samples compared to exploration samples as proach described above is used when several
the turn-around time between dispatching the flitches are sampled using inclined drilling. The
samples and receiving the results needs to be' geological interpretation is carried out on sections which are then sliced to produce flitch
much higher.

228

J,

G.

DAVIS

I
m

~-~
N

~1~

=\~,

o*1~

~
o

. .

J~
o

o
O

_o

~ o O O O ~

'~.

8.

~o

,;

,~

.-&
o

r-..:

8
~

~.

~/"~
t<

OPEN PIT GRADE CONTROL

plans. The ore outlines can the be modified to

produce the eventual ore extraction pattern.

Ore extraction
Hydraulic excavators have replaced front end
loaders for selective mining of the open pit gold
mines in Australia and the rationale behind this
move has been described by Pinniger (1984)
based on experiences at WMC's Sand King deposit near Kalgoorlie in Western Australia. The
most important consideration is the reduction
of dilution of the ore by surrounding mullock
(waste) when zones as thin as 0.5 m are being
mined separately.
The ore excavation needs to be closely supervised from the grade control point of view for
two reasons. Firstly, the trucks being loaded
must be dispatched to the correct stockpile
destination. Secondly, if it is possible to visually
discern the ore-waste contacts then the ore loss
and dilution can be minimized by detailed
instructions to the excavator operator and, in
theory, selective mining can be carried out on
excavator bucket sized units. Inspection of the
faces exposed as the ore is removed also allows
the geologist to ear mark material, that appears
to be mineralized but which may have been
missed by the in-situ sampling, for truck dumping and re-sampling as described above.

229

Drill and blast situation

In selective mining operations two factors which
become important in blast designs are: (i) the
minimization of the intermixing of ore and waste
and, (ii) the limitation of excessive movement of
blasted material.
If these factors are not controlled then the
preliminary grade control sample results will no
longer bare any relation to the material being
mined and the selectivity of the mining method is
lost. The potential for ore loss and dilution at
the time of mining is therefore greater when
blasting is required.
The two most widely used blasting techniques
in Australian open pit gold mines are bench
blasting, where there is a vertical free face to the
blast, and paddock blasting, where only the top
surface of the flitch is exposed at the time of
blasting. Little & Van Rooyen (1988) conducted
a survey of blasting techniques employed on
the Eastern Goldfields of Western Australia and
further divided the two techniques according
to whether ore and waste are blasted simultaneously or separately. The most common
approach is to bench blast ore and waste together and then to mark out the ore zones over
the broken ground. The excavator then removes
the ore and waste working across the strike of
the lode and away from the free face.

Reconciliation and recoverable reserves

Free dig situation

Before extraction begins from a flitch, the ore
zones are marked out on the pit floor using
coloured tape or lime. The geologist may alter
the pattern shown on the dig plan according to
the presence of geological features observed in
the pit.
The ore is then dug using a hydraulic excavator fitted with a bucket with a capacity of about
2 bcm. The excavator works across the strike of
the ore body loading trucks which are either on
the flitch below ('bottom loading') or on the
same flitch (' top loading'). The top loading mode
of operation is preferable as the newly exposed
features of the lower flitch are not disturbed by
the machinery travelling over them. Spillage of
material onto the lower flitch is also minimized
in the top loading operations. It is normal in this
situation to remove the mineralized portions of
the flitch first and to then extract the waste on
either side using an increased flitch height and an
excavator which has a larger bucket size. If the
material is soft enough, scrapers can be used to
remove the waste as this increases production
rates.

An important result of the daily grade control

procedures is the accumulation of estimated pit
production in terms of tonnages and grades of
each stockpiled ore category. These estimates
can be compared with the reported mill production figures and recoverable reserve inventory on
a monthly basis. This exercise highlights problems with the grade control methods and allows
the remaining recoverable reserves estimates to
be updated to take into account any consistent
discrepancies between the predicted and actual
bench productions.
The grade and tonnage produced from each
area of the pit worked during a reconciliation
period is estimated by collating information
from mill feed sampling and weightometer readings, stockpile tonnage estimates, pit surveys and
records of the number of truck loads taken from
each area. There are many possible sources of
error in this procedure and these are exacerbated
when ore from several sources is treated by the
same mill, however, these figures are considered
to be the most accurate assessment of the true
tonnage and grade of each bench.
The predicted, mineable tonnage and grade

230

J . G . DAVIS

from each bench can be estimated from the

bench grade control plans. The tonnage of each
stockpile category is calculated from the areas of
the blocks marked out, the flitch height and the
estimated in-situ density of the ore. In most
cases, the grades assigned to each block are
based on an arithmetic average of all the samples
within it. The main reasons for differences
between these estimates and the actual production are the result of dilution and ore loss as
described previously (see Fig. 3). Other sources
of errors are ascribed to the following.

Cutting of high grades

High grades are often cut to a nominal figure
such as 15g/t when they are included in the
calculations. These nominal figures may vary
throughout a deposit taking into account local
fluctuations in the assay distribution. The actual
figure chosen is sometimes purely arbitrary but
techniques based on the cumulative frequency
curve of the assay results are common.

Specific gravity
The specific gravity of the oxidized ore can vary
significantly, in particular there is normally an
increase in ore density downwards through a
weathered deposit. Allowances have to be made
for changes in specific gravity during the mine's
lifetime and these can have a marked effect on
the revised tonnage estimates of the remaining
reserves. In a similar manner, the moisture content of the clay material in the weathered zone
can vary significantly and has an impact on the
reconciliation calculations.
Additional errors due to sampling and analytical problems, the accuracy of stockpile volume estimation, and mined ore and waste being
dispatched to the wrong destinations, also lead
to discrepancies between mill and production
estimates. The exact source of the errors is difficult to locate and many mines therefore apply a
'mine call' factor to convert production estimates into 'diluted tonnes and grade'. These
mine call factors are derived empirically and are
used to incorporate all the errors described
above.

Computer applications
The role of the computer in Australian open pit
gold mining has become increasingly important
in recent years. At almost all mines a computer is
used to produce bench grade control plans such
as the one shown in Fig. 7. The extent to which a
computer is used thereafter in the grade control

programme varies from mine to mine but

manual methods are still very much in evidence.
Grade control sampling at a mine can produce
several hundred samples per day. The compilation of bench grade control plans and sections
from the assay results of these samples is a task
which readily lends itself to computerization. In
many cases, the sample grades can be transmitted directly to a mine site computer from the
assay laboratories using a modem link. Any one
of a number of propriety mining software packages may then be used to store the assays in a
database and generate the grade control plans
and sections. This speeds up the data processing
and reduces data entry errors.
The geologist's interpretation of ore outlines
are usually manually drawn onto hard copies of
the bench plans and these outlines can then be
digitized and stored in a database for reconciliation purposes. The co-ordinates of the points
forming the ore boundary outlines can also be
output from the computer in a form which
allows the pit surveyor to mark out their locations on the flitch.
The ability to produce a contoured plan of the
assay grades, onto which the original blast-hole
or Ditch Witch sample grades and geological
features can be overlain, is another feature made
available through the use of computers. The
contours can be produced by kriging or by triangulation though neither of these methods is
widely used at Australian mines. Nor is geostatistics routinely used in estimating bench
grades for reconciliation procedures.
The extra time involved in producing bench
plans which show kriged grade contours partly
explains the reluctance to adopt geostatistical
techniques at Australian gold mines. The erratic
nature of the gold grade distribution, combined
with a high co-efficient of variation and nugget
effect, also cause problems in producing meaningful semi-variograms on which the kriging
parameters are based. Furthermore, mine personnel are often unfamiliar with the techniques
and are wary of causing unwanted smoothing of
grades across ore zone boundaries.
Despite these points, there have been many
trials involving kriging in an attempt to improve
the reconciliation between mill production and
bench grade estimates (Snowden & D o w n e s
1988; Thomas & Snowden 1990).
The application of an indicator-type approach
has been investigated at the University of Wales
(Davis 1992). Data from a skarn-hosted gold
deposit was analysed and used to estimate
the grade and mineralized proportions of predefined, irregularly shaped ore blocks. This provides an indication of the likely internal dilution

OPEN PIT GRADE CONTROL

of individual blocks which can be compared to
production estimates. A computer program
which can combine this method with an interactive grade control bench plan editor has been
developed.

Operational aspects of grade control

Time is the most important factor in the grade
control programme. The operations are often
carried out on a hand-to-mouth basis and the
time between marking out a bench for sampling
and mining it may be only a few days. The main
delay is normally due to the analysis of samples
at commercial laboratories and the turn-around
time is therefore limited to between 18 and 24
hours.
Pressure for production rates to be increased
can cause a breakdown in the selectivity of the
mining method. This is a particular problem
when the mining is being carried out by earth
moving contractors rather than by the mining
company itself. Similar problems apply to the
sampling procedures which must be carried out
thoroughly if they are to be done accurately.
Gold mining companies are notoriously reticent about releasing figures on mining costs but
assaying costs of around AS10, combined with
labour costs of samplers and geologists, would
indicate the total cost of the grade control programme to be around one dollar per tonne of ore
mined. The costs in free-dig and drill and blast
situations vary but the overall cost is small
compared to the potential increased profit if the
grade control is effective. A figure of A$12.30
per ounce of gold produced has been quoted for

231

the annual cost of grade control in 1988 at New

Celebration Mine (Sullivan 1989).

Summary
Stringent grade control programmes are necessary to profitably mine many Australian open pit
gold mines. A flexible approach needs to be
taken to sampling procedures as the nature of
the gold distribution not only varies from mine
to mine, but often within the individual deposits.
The time scales involved in the running of the
mining operations requires cost effective processing of the sample results to be carried out
swiftly using computer technology where possible. However, only when the geological input
into the grade control strategy is maximized will
the results of the mill feed be optimised, that is,
dilution and ore loss will be minimized.
The ability to match closely spaced sample
information with observed geological features
exposed in the open pits is not only important
for mill reconciliation calculations but also helps
to build up a knowledge of the deposits which
can be incorporated into resource estimates at
similar locations. Grade control work is therefore seen as a continuation of the mineral
resources evaluation process and not merely a
guide to the short term extraction of ore and
waste.
The author would like to thank the many geologists
who provided the opportunity to visit the mines
described in this paper. The support of BHP Gold
Ltd, University of Wales College of Cardiff and the
NERC is also appreciated.

References
ARCHER, W. L. 1987. The development of the ditch
witch trenching machine as a grade control sampling tool. In: Equipment in the Minerals Indus-

COLLINGS,M., LENARTOWICZ,D. & ALEXANDER,K. R.

1988. Dominion Mining Limited--Cosmo Howley Gold Mine. In." Gold Forum on Technology

try." Exploration, Mining and Processing Conference, Kalgoorlie, WA October 1987. The
AuslMM Kalgoorlie Branch. AuslMM Symposium Series, 54, 21-23.

and Practices, World Gold 89. Proceedings of 1st

Joint International Meeting Between S M E and
AusIMM, November 1989. Reno, Nevada. 214-

BET'rENAY, L. F. & SHAW, W. J. 1990. Tabulation

of grade control techniques in representative
Australian mining operations. In." Strategies For
Grade Control, Australian Institute of Geoscientists Bulletin, |0, 96-102.

CARRAS,S. N. 1986. Concepts for calculating recoverable reserves for selective mining in open pit gold
operations. In: Selective Open Pit Gold Mining,
AusIMM Perth Branch Seminar, May 1986,
1-32.
- - 1987. Gold sampling--the importance of getting
it right. In: Meaningful Sampling in Gold Exploration. Australian Institute of Geoscientists Bulletin, 7, 1-25.

221.
DAVIS, J. G. 1992. The evaluation of Australian open pit
gold mines. PhD thesis, University of Wales.
DEVLIN, S. P. 1990. Grade distribution and grade
control procedures, Great Lady Gold Deposit,
Eastern Goldfields, WA. In: The AuslMM Mine
Geologists' Conference, Mount Isa, Queensland,
October 1990, 93-97.

GROVES, D. I., BARLEY,M. E. & Ho, S. E. 1988. The

nature, genesis and tectonic setting of mesothermal gold mineralisation in the Yilgarn Block,
Western Australia. In: Bicentennial Gold 88, Melbourne, Victoria, Australia, May 1988. Geological
Society of Australia.
LITTLE, T. N. & VAN ROOYEN, F. 1988. The current

232

J.G.

state of the art of grade control blasting in the

eastern goldfields. In: Explosives in Mining
Workshop, Melbourne, Victoria, November 1988.
AuslMM Symposium Series, 61, 87-95.
MANN, A. W. 1984. Redistribution of gold in the
oxidised zone of some Western Australian deposits. In: Gold Mining, Metallurgy and Geology,
The AuslMM Perth and Kalgoorlie Branches,
Regional Conference, October 1984, 447-458.
PISNIGER, H. R. 1984. The successful application of
hydraulic excavators for the selective mining of
auriferous open cut orebodies--with a little help
from the computer. In." Gold Mining, Metallurgy
and Geology, The AusIMM Perth and Kalgoorlie
Branches, Regional Conference, October 1984,
333-337.
R~CHARDS, A. 1990. Grade Control Practices at New
Celebration Gold Mine. In: Strategies For Grade
Control, Australian Institute of Geoscientists
Bulletin 10, 9-25.
ROYLE, A. G. 1992. A partial review of geostatistics.
This volume.
SAWYER, J. B. 1992. Assaying in resource evaluation
(The need for a clear and open mind). This
volume.
SCHWANS, P. B. 1987. The application of ditch witch
sampling in oxidised open cut gold mines. In:
Equipment in the Minerals Industry." Exploration,
Mining and Processing Conference, Kalgoorlie,
WA, October 1987. The AusIMM Kalgoorlie
Branch. AuslMM Symposium Series; 54, 25-31.

DAVIS
SHAW, W. J. 1990a. Grade control for gold mining at
Mt Magnet, Reedy, Rothsay and Youanmi in the
Murchison District, WA. In: The AuslMM Mine
Geologists' Conference, Mount Isa, Queensland,
October 1990, 83-92.
1990b. Management of grade control and mining
geology--an overview. In: Strategies For Grade
Control. Australian Institute of Geoscientists
Bulletin, 10, 1-7.
SNOWDEN, V. DOWSES, P. 1988. Geostatistics at
Marvel Loch gold mine; improved mine planning
and grade control. In: Bicentennial Gold 88, Melbourne, Victoria, Australia, May 1988. Abstracts
Geological Society of Australia, 23, 573-574.
SULLIVAN,B. W. 1989. Open pit planning and selective
mining practices at the New Celebration Gold
Mine. In: Gold Forum on Technology and Practices, World Gold 89. Proceedings of 1st Joint
International Meeting Between SME and
AusIMM, November 1989. Reno, Nevada, 139146.
THOMAS, M. & SNOWDEN,V. 1990. Improving reconciliation and grade control by statistical and geostatistical analysis. In: Strategies For Grade Control. Australian Institute of Geoscientists Bulletin,
10, 1-7.
WATCHORN, R. 1990. Open pit mapping aspects of
grade control: advantages and disadvantages. In:
Strategies For Grade Control. Australian Institute
of Geoscientists Bulletin, 10, 27-30.