HParker (Reconciliation Principles For The Mining Industry) (2014) .PDF Shortcut

Download as pdf or txt
Download as pdf or txt
You are on page 1of 17
At a glance
Powered by AI
Reconciliation involves comparing independent data sources such as exploration, production sampling, and plant data to measure accuracy and identify sources of error.

Reconciliation factors (F1, F2, F3) relate different data sources and enable comparisons of tonnage, grade and metal content estimated vs actual. F1 compares short term model to reserves, F2 compares mill to production, F3 compares mine to reserves.

Sources of error include inaccuracy in resource/reserve estimates and inefficiency in the mining process like inaccurate orebody knowledge and assumptions in selectivity.

Reconciliation principles for the mining industry

H. M. Parker*
Reconciliation involves the collection of tonnage, grade (quality) and contained metal (product)
data from disparate and hopefully independent sources. Examples are exploration data,
production sampling data from blast holes or draw points, and process plant data. These data
may be compared by means of ratios (factors). The F1 factor usually relates short term (ore
control) model tonnages, grades and metal content to ore reserves depleted. The F2 factor
usually relates received at mill (measured by the mill) tonnages, grades and metal content to
delivered to mill production tonnages, grade and metal content. The F3 factor is F16F2 and
enables a comparison of a mine’s (measured by mine) ability to recover the tonnage, grade and
metal content estimated in ore reserves. The F1 factor measures the accuracy of orebody
knowledge in the ore reserves to the demarcation of ore and waste by ore control (short term
model). The F1 factor may be used to check and calibrate the selectivity of mineral resource
models and/or planned dilution assumed in transfer from mineral resources to ore reserves. The
F2 factor enables a check on unplanned dilution entering the ore stream between ore control and
the mill. By using the factors it is possible to calculate a monetary value on improvements in the
accuracy of orebody knowledge, selectivity and the effects of dilution and ore loss. Reconciliation
should be an implicit part of the mining process, and reconciliation targets should be a key
performance indicator for well run mines.
Keywords: Reconciliation, Tonnage, Grade, Contained metal data, Factors, Ratios

Introduction reconciliation between mine and mill would be ¡10%. Even


so, it is useful to examine the impact of such seemingly small
This paper is an extension to Parker (2006) and builds estimation errors on the cash flows and net present values
on over 25 years of practical reconciliation experience (NPVs) of an operation. This is shown in Table 2.
gained in copper, gold, nickel and iron mining opera- The change in cash flow is three times that in grade;
tions located throughout the world. Reconciliation is a the change in NPV is six times that in grade. A small
key process that allows determination of the ability of a negative error can cause a mine to miss its cash flow
mining operation to produce the tonnage, grade and targets. If these persist, unfavourable press releases must
contained metal that were estimated in the ore reserve. be issued to the marketplace. For longer term planning,
This paper discusses the principles of reconciliation and mining companies weigh the NPVs of various alter-
the commonly encountered sources of error in reserve, natives when deciding to make acquisitions or determin-
ore control and mill head estimates by comparing the ing the sequence in which they will develop their assets.
data collected by means of ratios or factors. Before Errors may result in incorrect decisions, and may even
developing the principles, it is necessary to see why result in premature mine closure.
reconciliation is important. A simple example is shown To understand and avoid sources of error, it is useful
in Table 1, which shows a production schedule and cash to gather reconciliation data at various stages of the
flow statement for a copper mine, where the reserves are mining process. To be valid, the reconciliation data
estimated to grade 0?6% copper. taken for each stage should be independent, and
One of the goals of ore reserve estimation is to make an generally, both tonnage and grade should be estimated
unbiased estimate of the tonnage and grade. This is seldom or measured. The collected data are analysed in terms of
true in practice. Errors will occur, and these will be related ratios, herein called factors.
to inaccurate orebody knowledge at the time of ore reserve
estimation and to the assumptions made as to the accuracy
with which ore is selected from waste in the ore control
Sources of error
process. It is generally accepted that in a base metals mine, In most mines there are two main sources of error that
a good annual reconciliation between mine and mill would will be considered here:
be ¡5%, and for a precious metals mine, a good annual (i) inaccuracy in estimation of mineral resources
and/or ore reserves (long range model)
(ii) inefficiency in the mining process to segregate ore
AMEC, 780 Vista Blvd., Suite 100, Sparks, NV 89434, USA and waste as planned (short range model) by the
*Corresponding author, email [email protected] ore control staff.

ß 2012 The Australasian Institute of Mining and Metallurgy


Published by Maney on behalf of the Institute of Materials, Minerals and Mining and The AusIMM
Received 23 February 2012; accepted 8 May 2012
160 DOI 10.1179/1743286312Y.0000000007 Mining Technology 2012 VOL 121 NO 3
Table 1 Production schedule and cash flow statement for copper mine
Input variable
Mining ore/$/t 2.00
Mining waste/$/t 2.00
Milling/$/t 5.00
General and 4.00
admin./$/t
Copper price/$/lb 2.50
Conc. Grade/%Cu 29.0
Conc. recovery/% 90
Freight/$/t Conc. 50
Smelting/$/t Conc. 80
Refining/$/lb 0.08
Payable Cu/% 95
Pounds/t 2204.62
Year 22 21 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Production schedule
Ore mined/Mt 10 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20
Waste mined/Mt 10 20 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40
Ore grade/%Cu 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60
Concentrate/Mt 0.186 0.372 0.372 0.372 0.372 0.372 0.372 0.372 0.372 0.372 0.372 0.372 0.372 0.372 0.372 0.372
Payable Cu/Mt 0.051 0.103 0.103 0.103 0.103 0.103 0.103 0.103 0.103 0.103 0.103 0.103 0.103 0.103 0.103 0.103
Cash flow statement/M$
Revenues 282.7 565.5 565.5 565.5 565.5 565.5 565.5 565.5 565.5 565.5 565.5 565.5 565.5 565.5 565.5 565.5
Ore mining 220.0 240.0 240.0 240.0 240.0 240.0 240.0 240.0 240.0 240.0 240.0 240.0 240.0 240.0 240.0 240.0
Waste mining 240.0 280.0 280.0 280.0 280.0 280.0 280.0 280.0 280.0 280.0 280.0 280.0 280.0 280.0 280.0 280.0
Milling 250.0 2100.0 2100.0 2100.0 2100.0 2100.0 2100.0 2100.0 2100.0 2100.0 2100.0 2100.0 2100.0 2100.0 2100.0 2100.0
GzA 240.0 280.0 280.0 280.0 280.0 280.0 280.0 280.0 280.0 280.0 280.0 280.0 280.0 280.0 280.0 280.0
Freight 29.3 218.6 218.6 218.6 218.6 218.6 218.6 218.6 218.6 218.6 218.6 218.6 218.6 218.6 218.6 218.6
Parker

Smelting 214.9 229.8 229.8 229.8 229.8 229.8 229.8 229.8 229.8 229.8 229.8 229.8 229.8 229.8 229.8 229.8
Refining 29.0 218.1 218.1 218.1 218.1 218.1 218.1 218.1 218.1 218.1 218.1 218.1 218.1 218.1 218.1 218.1
Capex 2200.02200.02200.02100.0 230.0 230.0 230.0 230.0 230.0 230.0 230.0 230.0 230.0 230.0 230.0 230.0 230.0 230.0 0.0
Pretax cash flow 2200.02200.02200.0 20.5 169.0 169.0 169.0 169.0 169.0 169.0 169.0 169.0 169.0 169.0 169.0 169.0 169.0 169.0 199.0

Mining Technology
Cumulative 2200.02400.02600.02600.5 2431.5 2262.6 293.6 75.4 244.4 413.3 582.3 751.3 920.3 1089.2 1258.2 1427.2 1596.2 1765.1 1964.1
cash flow
Net present value 554.3

2012
(NPV) at 8%

VOL
121
NO
3
Reconciliation principles for the mining industry

161
Parker Reconciliation principles for the mining industry

Table 2 Sensitivity of cash flow, net present value and payback to changes in grade (0?60%Cu is base case)

%Cu %Change Cash flow year 2/M$ %Change NPV at 8% %Change Payback years

0.66 10 218.9 30 886.5 60 4


0.63 5 193.9 15 720.4 30 4
0.60 0 169.0 0 554.3 0 5
0.57 25 144.0 215 388.1 230 6
0.54 210 119.1 230 222.0 260 7

The inaccuracy in estimation of mineral resources and/ layouts. Where caving methods are used, drawpoint
or ore reserves is determined by comparing depletions samples and observations of lithology are used to decide
from the long range model to the short range model. when to shut off drawpoints.
Inefficiency in the mining process is determined by The ratios between tonnage, grade and metal in the
comparing received at mill tonnages, grade and metal long range and short range models are used to develop
content based on mill tonnage measurements and head reconciliation factors, known as F1*, as F1 (tonnes), F1
assays versus delivered to mill tonnages, grade and (grade) and F1 (metal)
metal content based on the short range (ore control)
model. short range model depletions
F1~
There are other sources of error listed below that can long range model depletions
be important but will not be considered: Where a deposit has more than one grade variable (such
(i) failure to mine within planned areas
as Cu and Au), then there will be multiple F1 factors for
(ii) failure to mine at the planned cut-off
grade [F1 (%Cu), F1 (Au g/t)] and for metal [F1
(iii) plant performance.
(contained Cu), F1 (contained Au)]. In some deposits
There may be very good mitigating circumstances, such
F1 factors are developed for deleterious elements.
as change in markets or physical impediments (slides,
Sometimes a deposit will produce several products. In
stope/drawpoint collapse), but in the long run mines that
these cases, F1 factors are constructed for each product.
fail to follow the plan tend to find themselves with a
A high value for an F1 factor indicates conservatism
shortfall in stripping or remnants in ore reserves that
in the long range model; conversely a low value for an
have to be written off.
F1 factor indicates optimism in the long range model.
Reconciliation within treatment plants has been
discussed at length by Morrison (2008). Table 3 describes sources of error and likely resultant F1
factors.
Reconciliation between long range model and As can be seen in Table 3, the sources of error can be
short range model ambiguous. For example, a low F1 (tonnes) could be
just as likely to be related to an over smoothed long range
The long range model is constructed using geological
model as to conservative dig lines creating remnants. In
interpretations based on exploration and/or delineation
general, common sense should prevail; where the drill
drill holes. Assays from these holes are composited
and used to estimate grades in blocks. The result is a hole spacing is too wide to identify discrete ore zones, the
resource model. In massive deposits with broad trends to long range model will be inaccurate. More drilling is
grade, a pit is developed using mining software and needed to support the long range model. This may not
refined and scheduled. The cut-off grade may be changed occur everywhere in the deposit, but only in specific areas
over the course of the schedule using the theory of Lane (e.g. lithological contacts, structures). In addition, con-
(1988). In such deposits, the ore reserve is often the sideration should be given to whether the mine is
tonnage of measured and indicated mineral resources undergoing stress, as for instance recovering from a
within a pit. geotechnical event such as a slide after which there are
In deposits with sharp decreases in grade at ore/waste fewer than normal working areas available.
contacts, the resource model may be adjusted to allow
for dilution and less frequently ore loss at the contact. Reconciliation between delivered to mill and
These adjustments may take the form of aggregating received at mill
small blocks into large blocks or adjusting the grades of Most mines have transit stockpiles where ore is stored
ore blocks by a given amount at their faces with waste temporarily. Therefore, material leaving the pit and
blocks. In underground mines it is common to design being depleted from the short range model can be
stopes and to apply dilution and recovery factors. Where delivered to a stockpile or to the mill (at its short range
block caving will be used, complex software such as PC- model grade). Generally, stockpiles are assigned the
BC is often used to make adjustments for various average grade of all the increments added to them. If the
amounts of mixing in the cave profile. stockpiles are reclaimed and rebuilt fairly regularly (such
The short range model is commonly constructed using as to allow for a bad weather season with low pro-
samples from closely spaced drill holes and/or mapping ductivity from the pit), then this should not significantly
information. This information is used to delineate ore affect reconciliation much, particularly if the quantities
and waste in mineable shapes. In an open pit mine the represented in the numerators and denominators of the
shapes are staked in the field, sometimes with adjust- factors have been aggregated over quarters or years.
ment for blast heave (Yennamani et al., 2011; La Rosa Reconciliation in the cases where old/large stockpiles are
and Thornton, 2011). In an underground mine, cross-cuts,
*The F1, F2 and F3 factors and the terms ‘delivered to mill’ and ‘received
channel/chip samples and short drill holes (commonly at mill’ come from the Nchanga Open Pits survey department, Chingola,
drilled on 15 m spacings) are used to make the final stope Zambia.

162 Mining Technology 2012 VOL 121 NO 3


Parker Reconciliation principles for the mining industry

reclaimed or where a large percentage of production is tonnage can get out of calibration; in addition moisture is
related to stockpile reclaim can be problematic. normally backed out of the tonnage.
At larger mines there is often either no mill head The ratios between tonnage, grade and metal content
sampler or a recirculation/mixing of materials within the received at mill to delivered to mill are used to develop
mill that prevents a head sample from being taken. This reconciliation factors, known as F2, as F2 (tonnes), F2
means that the head grade is calculated from the tailings (grade) and F2 (metal)
and concentrate tonnages and grades. Concentrate ton-
received at mill
nages and grades are usually accurately measured. The F2~
tonnage and grade of the tailings stream are sometimes not delivered to mill
reliable, which can affect the calculated head grade. The Delivered to mill is the combination of short term model
weighing devices used to measure the received at mill depletions expit direct to mill and stockpile reclaim at

Table 3 Sources of error and likely F1 factors

F1 F1 F1 Comments/
Source of error (tonnes) (grade) (metal) remedial action

Mainly related to long range model


Conservative grade shells, .1 ,1 .1 Usually the tonnage effect is greater than the
ore bearing lithologies grade effect, giving a positive F1 (metal). At early
stages (wide spaced drilling), conservatism may be
justified, particularly for inferred resources. Projection
of grade shells and favourable lithologies on two sets
of orthogonal sections and then to plan may help;
Also, recognition of contacts, structural, lithological
or alteration controls may help in projecting further in
sparsely drilled areas. The use of grade shells may
cause distortions in the grade–tonnage curve. These
can be mitigated by sample sharing at boundaries.
In the end, more drilling is likely needed to upgrade
orebody knowledge. Will increase ore tonnage in the
long range model.
Over smoothed long ,1 .1 About This can be a very serious error when processing costs
range model same will be high. The long range model incorporates too much
dilution, which ore control can segregate, with very
favourable financial consequences. A stop gap solution
is to interpolate grade using fewer samples or a high
power inverse distance method. This can have a bad
knock-on effect of local conditionally biased estimates.
It is better to use deterministic or probabilistic domaining
or a Uniform Conditioning/MIK method. However,
the best solution is to tighten the drill spacing
Under-smoothed .1 ,1 About Domains and grade shells are too selective. Consider
long range model same some mixing of samples across domain boundaries
with outlier restriction or topcutting to control over
projection of grades.
Grade effects About .1 .1 Where grade distributions are very skewed, wide
same spaced drilling will undersample the high end of the
distribution. In diamond deposits, large stones may be
broken. If core recovery is poor, may underestimate
grade, examples are chalcocite, molybdenite and
gold. Plucking of soft materials may also result in
underestimation of grade.
Reverse circulation About .1, .1, In dry RC, fines may not be recovered; this may affect
drilling same ,1 ,1 grade. For wet RC, fines may be washed out; gold may
be concentrated (placering). Weak high grade zones
may contaminate samples from above. Drilling
practices must be improved with mud conditioning,
flocculants in buckets, better splitting,
or switching to drilling of core (deep holes).
Mainly related to short range model
Sloppy dig line layout .1 ,1, .1 About Effective SMU is larger than plan. Very common for
same low grade dump leach or stockpile categories with a
few erratic high grade values. Better ore control
sampling (such as every blast hole versus one in
six) can help. Adding short delineation holes may
help. Improve quality of face or blasthole sampling
(increase size of sample taken). Decrease block size
Biased sampling About same .1 .1 Over sampling of fines in blast hole or grab samples
can cause a bias
Leaving remnants ,1 .1 ,1 Purposely conservative final stope design in times
of economic stress

Mining Technology 2012 VOL 121 NO 3 163


Parker Reconciliation principles for the mining industry

average grade based on short term model increments


placed on the pile.
Where a deposit has more than one grade variable
(such as Cu, Au), then there will be multiple F2 factors
for grade [F2 (%Cu), F2 (Au g/t)] and for metal con-
tent [F2 (contained Cu), F2 (contained Au)]. In some
deposits F2 factors are developed for deleterious elements.
Sometimes a deposit will produce several products. In
these cases, F2 factors are constructed for each product.
A high value for F2 will indicate conservatism in the
short term model. A low value for F2 will indicate
optimism in the short term model.
In the case of underground mines, stockpiles are not
often used, but remuck bays provide temporary storage
and an opportunity for mixing of ore and waste.
Table 4 describes sources of error and likely F2
factors.
With the advent of GPS control and automated 1 Summary relationships between factors
dispatch systems, F2 factors are now generally close to
unity in most open pit mines. There are still problems
with mixing of ore and waste in underground mines, short range model depletions
F3~ |
particularly where development is behind schedule. It is long range model depletions
just too easy to stash waste development tonnage into
received at mill
the ore delivery system to the mill. ~F1|F2
Sometimes it is preferable to overbreak contacts to delivered to mill
ensure against ore loss; however, this should be planned Over a period of time (annual) short range model
and thus would give a high F1 (tonnes) factor and low depletions will cancel with delivered to mill.
F1 (grade) factor. Properly done, there would be no One of the advantages of the F3 metric is that it
effect of this practice on F2 factors. removes the effect of ore control sampling (grade) bias,
which often afflicts ore control.
Reconciliation between mill and long range Figure 1 shows a summary depiction of the factors
model and their relationships.
The long range model (sometimes after adjustment for
planned ore loss and dilution) is the basis for the life of Remarks
mine plan and through it cash flow forecasts. An F3 It is becoming more frequent to use the F1, F2 and F3
factor can be created which allows for the effects of frameworks at the feasibility stage to test or adjust the
stockpiles long range model to reflect the intended degree of

Table 4 Sources of error and likely F2 factors

Source of error F2 (tonnes) F2 (grade) F2 (metal) Comments/remedial action

Mainly related to short range model


Biased sampling About same ,1 ,1 Over sampling of fines in blast hole
or grab samples can cause a bias
Truck allocation Either high/low .1, ,1 .1, ,1 Miscount on trucks delivering to mill, stockpiles
Mainly related to poor mining practice
Overbreak .1 ,1 About same Mining beyond dig lines; following
colour changes in muck unrelated to
ore control boundaries; bonus scheme
should be based on metal, not tonnes broken
Blast heave, unplanned About same ,1 ,1 Blast ore and waste in opposite directions;
mixing of ore and waste separate ore and waste passes, remuck
bays; keep waste out of crusher; use
modern dispatch system
Pulling too fast in block ,1 ,1 ,1 Piping at drawpoints; drawpoints close
cave mines early due to low grade factors
Mainly related to received at mill
Bad tonnage Either high/low No effect No effect Belt scales not calibrated; moisture is
guestimated. Calibrate scales every two
weeks; obtain weekly moisture samples;
determine seasonal corrections
Bad grade No effect ,1, .1 ,1, .1 Usually F2 (grade) ,1; need better tailings
tonnage and grade measurement and samplers
Multiple products
,1, .1 ,1, .1 ,1, .1 Mill campaigns may not coincide with ore
types as segregated in pit; better
communication can help

164 Mining Technology 2012 VOL 121 NO 3


Parker Reconciliation principles for the mining industry

2 Blast blocks (black) and blast hole (coloured) grades (%Cu)

selectivity. Conditional simulation was used by Edward was reflected in the short term model as measured by F1’,
Isaaks as early as 1991 to adjust an over smoothed and whether the 20620615 m SMU demonstrated the
kriged model for the Lihir gold deposit, located in same degree of selectivity shown by ore control polygons
Papua New Guinea (Parker, 1992a). Table 5 shows an (diglines) as measured by F10. Table 6 shows the results
example of adjustment factors applied for various as follows:
selective mining unit (SMU) sizes. The corrections for (i) for copper only ores the F1’ factors show that
smoothing in the long range model are more severe as the long range model is slightly underestimating
the SMU size is decreased. tonnage and grade of blast blocks
This was repeated in 1994 for the Fort Knox Gold (ii) for copper–zinc ores the F1’ factors show that
deposit in Alaska. For both Lihir and Fort Knox, a very the long range model is underestimating tonnage
fine grid of grades was simulated at y3 m spacing, and and over estimating grade. The long range model
these were taken as ground truth. A subset was taken to is too selective
represent blast holes; ore and waste diglines were drawn (iii) for copper only ores the F10 factors show that
manually, and the ore control grade was estimated using ore control polygons have similar tonnage and
kriging. The resultant ground truth grades within grade to the blast blocks. The ore control
diglines were used to formulate received at mill grades. polygons have a level of selectivity implicit in
At Quebrada Blanca (Iquique, Chile), conditional a 20620615 m SMU
simulation was used to determine the impact of over- (iv) for copper–zinc ores the F10 factors show that the
breaking the leached cap contact with supergene ore control polygons have much higher tonnage
enriched ore to minimise ore loss. at the same copper, but much higher zinc grades.
It is sometimes useful to break the factors down This was later found to be related to over
further. For example at Antamina Mine in Peru, a factor projection of high grades in the short range model.
F1’ was developed that compared the distribution of
blast blocks with the long range model blocks. Blast At BHP Billiton, the F3 factor is defined slightly
blocks had the same dimensions as the long range differently. The contained metal/coal in saleable product
model blocks (20620615 m). The ore control model (beneficiated iron ore, marketable coal, or metal con-
(565615 m blocks) was aggregated to blocks with tained in concentrates) for a time period is the nume-
dimensions of 20620615 m (see Fig. 2). rator. The denominator is the forecast metal/coal
This enabled determination as to whether the selectiv- recovered from the life of mine plan adjusted for
ity implicit in the long range model (20620615 m SMU) stockpile additions and depletions.
Therefore, BHP Billiton is taking into account
treatment plant performance as well. BHP Billiton also
Table 5 Adjustment factors for Lihir (Parker, 1992a) reports F1 and F2 factors as defined herein.
Sometimes the F1 factors are referred to as block
SMU F3
factors which are related to the pregeostatistical era
Size/m Volume/m3 Tonnes Grade Metal
when factors were applied to polygonal and other simple
ore reserve estimates. As an example at Jerritt Canyon
66662 72 0.78 1.18 0.92 Nevada in the 1980s, a nearest neighbour estimate was
66664 144 0.82 1.12 0.91 made of gold grades. For production scheduling, grades
126563 180 0.84 1.09 0.91 above cut-off were multiplied by 0?92, and tonnages
96966 486 0.87 1.05 0.91
were multiplied by 1?05 (Parker, 1992b).

Mining Technology 2012 VOL 121 NO 3 165


Parker Reconciliation principles for the mining industry

Table 6 Detailed comparison of long range to short range models at Antamina Mine

Copper only ores

Mtonnes %Cu %Zn kt cont. Cu kt cont. Zn

2001–2004
Long term model* 66.8 1.49 0.25 993 165
Blast blocks{ 71.1 1.54 0.23 1096 164
F1’ factors 1.06 1.04 0.93 1.10 0.99

2003–2004{
Blast blocks 32.5 1.59 0.25 517 81
Ore control polygons 32.3 1.67 0.22 538 72
F10 factor 0.99 1.05 0.89 1.04 0.89
Copper–Zinc ores

Mtonnes %Cu %Zn kt cont. Cu kt cont. Zn

2001–2004
Long term model* 46.7 0.98 2.94 457 1375
Blast blocks 55.4 0.96 2.66 532 1474
F1’ factors 1.19 0.98 0.90 1.16 1.07

2003–2004{
Blast blocks 22.7 1.01 2.83 229 642
Ore control polygons 28.8 1.02 3.17 294 911
F10 factor 1.27 1.01 1.12 1.28 1.42
*Long range model built in 2005 after infill drilling programme.
{Blast hole model reblocked to 20620615 m blocks.
{Comparison of blast blocks to ore control polygons for 2001–2002 is not shown because oxidised material was sent to waste based
on pit mapping; this information was not preserved in blast blocks.

The F2 factors are also sometimes referred to as mine unacceptable to mine management. An investigation
call factors. These have variously covered a myriad of showed that more closely spaced drilling would be
issues such as mining recovery, unplanned dilution and required to delineate the high grade zones which were
ore loss. found to be discrete bodies within the deposit. The high
grade breccias were delineated as a separate unit for
Case study for Antamina Mine copper grade estimation. The high grade zinc zones were
found to occur near the marble/green exoskarn contact,
Antamina Mine is located in the Peruvian Andes and
but their variable width (20–40 m) was only able to be
extracts copper, zinc, molybdenum, silver and bismuth
domained through grade shells interpreted from detailed
from endoskarns and exoskarns developed in limestones
drilling.
surrounding a porphyry intrusive. Copper-only ores are
The mine staff and mine owners were in agreement
concentrated in endoskarns inboard from copper–zinc
that the drill hole spacing had to be reduced, but by
ores that are concentrated in exoskarns. Copper and
zinc concentrates are produced and shipped by slurry
pipeline to a port at Huarmey on the Pacific Ocean.
Molybdenum and minor lead concentrates are produced Table 7 Antamina Mine reconciliation between long range
as byproducts. Nominal ore production is y30 Mtpa. and short range models (Q1–3, 2003)

Reconciliation in 2003 Copper only ores


Table 7 shows reconciliation between long range and
short range models for the first to third quarters (Q1–3) Mtonnes %Cu %Zn kt cont. Cu kt cont. Zn
for 2003. In this case the long range model was produced
Long range 6.39 1.41 0.16 90.1 10.2
using feasibility-stage exploration drilling. model
The F1 factors are far from unity. For the copper only Short range 8.41 1.73 0.16 145.5 13.5
ores occurring in endoskarns, the 75–100 m spaced model
exploration drilling undersampled narrow (5–20 m F1 factors 1.32 1.23 1.00 1.61 1.32
wide) zones of high grade breccias (see Fig. 3). This
Copper–zinc ores
led to F1 (tonnes) of 1?31 and F1 (%Cu) of 1?22. For the
copper–zinc ores, the resource model based on 50–75 m Mtonnes %Cu %Zn kt cont. Cu kt cont. Zn
spaced exploration drill holes was smoothed, resulting in
F1 (tonnes) of 0?87 and F1 (%Zn) of 1?20 (see Fig. 4). In Long range 15.6 1.11 2.56 173.2 399.4
these ores F1 (%Cu) was lower (0?86), and reflects lower model
copper grades in the higher grade zinc areas. Short range 13.5 0.95 3.06 128.3 413.1
model
The reconciliation over the shorter time periods used
F1 factors 0.87 0.86 1.20 0.74 1.03
to plan mill campaigns and concentrate shipments was

166 Mining Technology 2012 VOL 121 NO 3


Parker Reconciliation principles for the mining industry

3 Antamina Mine 2003 long range Cu model (left) and short range model (right): red colour indicates .2%Cu, localised
in breccia zones; blocks (left) are 20620 m and 565 m (right)

how much? Conditional simulation could have been range models and the short range model (ground truth),
used for this purpose, but was rejected as being too time there was no noise introduced by using different sample
consuming; both multiple grade variables (Cu and Zn) types (such as drill cores versus auger samples of blast
that were required to be estimated, as well as hole cuttings).
lithological contacts. Instead, subsets of blast holes Figure 5 shows the results of three grids for zinc and
representing about 60 Mt were used to create pseudo the ground truth for comparison. The high grade zone
exploration grids, and these were in turn used to does not become well delineated until a 25625 m
estimate experimental long range models. The experi- spacing is achieved.
mental long range models were then compared with the Table 8 shows the change in bias as the drill spacing is
short range model estimated using all the blast holes reduced. The 1% cut-offs for copper and zinc represent
(and taken as ground truth). Since blast hole assays typical cut-offs used in 2003 by ore control for mill ore.
were used as the basis for both the experimental long The 2?5% cut-offs were designed to evaluate the high

4 Antamina Mine 2003 long range Zn model (left) and short range model (right): red and magenta colour indicates
.2?5%Zn, localised in structural zones; blocks (left) are 20620 m and 565 m (right)

Mining Technology 2012 VOL 121 NO 3 167


Parker Reconciliation principles for the mining industry

5 Zinc grade zones for three grids interpreted from experimental grids of blast holes: high grade zinc over 2?5% is red;
medium grade zinc (0?25–2?5%) is blue: a 75675 m spacing (high grade is under represented); b 50650 m spacing
(high grade is still under represented); c 25625 m spacing (high grade is fairly well represented); d ground truth
based on blast holes 767 m spacing

grade zones that contain about half the metal above cut- (ii) using a 2?5% copper cut-off, the F1 criterion for
off. An F1 factor of ,1?1 was deemed desirable to declare tonnage, grade and metal is achieved at a
indicated mineral resources at the 1% cut-off. To declare 50625 m spacing
measured mineral resources, it should be a goal to ensure (iii) using a 1% zinc cut-off, the F1 criterion for
that there are no biases at the high (2?5%) cut-off. tonnage and metal is not achieved for any
Examination of Table 8 shows: spacing. The F1 criterion is met at a 50625 m
(i) using a 1% copper cut-off, the F1 criterion for spacing for grade (%Zn)
tonnage, grade and metal is achieved at a (iv) using a 2?5% zinc cut-off, the F1 criterion is met
50650 m spacing for metal at the 25625 m spacing. Tonnage is

168 Mining Technology 2012 VOL 121 NO 3


Parker Reconciliation principles for the mining industry

Table 8 Antamina Mine F1 factor reconciliation between interest, it was decided that the greater uncertainty for
test long range and short range models (60 Mt) zinc could be accommodated within these drill spacings.
1% copper cut-off grade applied
Reconciliation after infill drilling
Drill spacing/m F1 (tonnes) F1 (%Cu) F1 (cont. Cu) Approximately 100 000 m of infill drilling was performed
in 2004. This brought nearly all the ore mined to the end of
75675 0.97 1.15 1.11 2004 to measured mineral resource status. Table 9 shows
50650 0.93 1.01 0.93
the reconciliation that was achieved during 2003 and 2004.
50625 0.88 1.02 0.90
25625 0.92 1.01 0.93 For this comparison a new long range model was
used; this model was constructed in 2005 after completion
2.5% copper cut-off grade applied of infill drilling. For copper only ores the reconciliation is
good. F1 (tonnes) and F1 (%Cu) are still higher than 1?0,
Drill spacing/m F1 (tonnes) F1 (%Cu) F1 (cont. Cu)
but only marginally. There has been great improvement
75675 2.04 1.14 2.33 compared to the results shown in Table 8 for previous
50650 1.11 0.96 1.06 estimates. The F2 factors are ,1?0. This likely indicates
50625 1.05 0.98 1.03 that some copper only ores were routed to copper–zinc
25625 1.03 1.02 1.04 campaigns by the mill, thus explaining F2 (tonnes) at
1% zinc cut-off grade applied 0?93. The mill head grade is lower than the short term
model depletions [F2 (%Cu)50?91]. This is likely evidence
Drill spacing/m F1 (tonnes) F1 (%Zn) F1 (cont. Zn) of high bias in blast hole sampling (meaning the auger
method used for ore control samples oversamples fines).
75675 1.52 1.35 2.04 The F3 factors for tonnage, %Cu and contained copper
50650 1.35 1.20 1.64
are near parity (0?99–1?01).
50625 1.43 1.06 1.52
25625 1.35 0.92 1.25 For copper–zinc ores, the F1 (tonnes) and F1 (%Zn)
are both greater than 1?0. The tonnage increase in short
2.5% zinc cut-off grade applied term model depletions compared to the long range
model is either related to overbreaking the marble/
Drill spacing/m F1 (tonnes) F1 (%Zn) F1 (cont. Zn)
exoskarn contact where high grade zinc occurs, or a
75675 3.70 0.69 2.56 conservative long range model (probable from Table 8).
50650 2.22 0.79 1.75 The former explanation is more likely. The received at
50625 1.37 1.00 1.37 mill grade is less than the short range model depletions
25625 1.18 0.88 1.04 (2?61 versus 3?21%Zn). Subsequent investigations
revealed the short range model was over projecting high
grade blast holes; this finding accounts for some of the
under estimated, and grade is over estimated. observed bias.
This is an artifact of the grade zoning process,
and is now ameliorated by sharing samples
Comments
across the high grade zone boundary. The reconciliation data were extremely useful:
(i) the F1 factors measured in 2003 identified a
In 2004 the drill spacing was reduced from 100 to 50 m to problem with the long range model
support indicated mineral resource declaration and from (ii) the blast hole data were used to demonstrate the
75 to 25–35 m to support measured mineral resource improvement to be gained in the accuracy of the
declaration. As copper is the primary metal of economic long range model by infill drilling

Table 9 Antamina Mine reconciliation for 2003–2004

Copper only ores

Mt %Cu %Zn kt cont. Cu kt cont. Zn

Long range model depletions 30.5 1.51 0.25 460.6 76.3


Short range model depletions 32.3 1.67 0.22 539.4 71.1
Received at mill 30.1 1.52 0.25 457.5 75.3
F1 factors 1.06 1.11 0.88 1.17 0.93
F2 factors 0.93 0.91 1.14 0.85 1.06
F3 factors 0.99 1.01 1.00 0.99 0.99
Copper–zinc ores

Mt %Cu %Zn kt cont. Cu kt cont. Zn

Long range model depletions 23.9 0.96 3.04 229.4 726.6


Short range model depletions 27.1 1.07 3.21 290.0 869.9
Received at mill 27.6 1.01 2.61 278.8 720.4
F1 Factors 1.13 1.11 1.06 1.26 1.20
F2 Factors 1.02 0.94 0.81 0.96 0.83
F3 Factors 1.15 1.05 0.86 1.21 0.99

Mining Technology 2012 VOL 121 NO 3 169


Parker Reconciliation principles for the mining industry

6 Histograms for composites and resource model

(iii) the F2 factors for 2003–2004 identified sam- tonnage; the average grade of the SMUs is always
pling biases in blast holes and over projection of higher than the average grade of the resource model.
high grade blast holes for zinc Table 10 shows the tonnages and average grades at a
(iv) the F3 factors measured for 2003–2004 show 1?5 g/t cut-off, with associated F1 factors. Table 11
that the tonnage, grade and metal for copper shows a comparison for various mining scenarios.
only ores forecast by the long range model are SMU2 would be considered the base case for a
achievable 10 Mtpa operation with 3 Mtpa of ore (10 000 tpd). It
(v) the F3 factors measured for 2003–2004 show can be seen that there is very little improvement in cash
that the zinc metal forecast by the long range flow for a more selective scenario (SMU1), and it might
model is achievable. However, this comes at a be impractical to be more selective in light of the need
higher tonnage and lower grade. Future long for more working room for smaller equipment. For
range models should consider allowing for SMU3 and SMU4 the effect of dilution and high
overbreak of the marble/exoskarn contact. treatment costs negates economies of scale in mining.
The cash flows for a plan based on the resource model
are about $40 million/year less than the base case.
Case study for hypothetical gold deposit The lesson should be clear. Selectivity matters,
Schofield (2001) has raised the issue that reconciliation particularly where treatment costs are high. An over
of the resource model and production does not go far smoothed model destroys value, as was eloquently
enough. Inappropriate resource models and/or ore shown by Zhang (1998). The solution suggested by
control practices can cause value to be lost. The case Zhang was multiple indicator kriging. Since then the
study presented below is typical of Carlin-type gold industry in the USA has turned more toward probabil-
deposits in the USA. It shows how using too smoothed a istic domaining of high grade zones.
resource model can destroy value. Figure 6 shows a
histogram for composites and a resource model. It is
possible to make a change of support transformation to
Case study for hypothetical iron ore
the composite distribution to obtain distributions of deposit
SMU grades (Journel and Huijbregts, 1978). The This case study was motivated by discussions with Geoff
discrete Gaussian method has been used for this case. Ballantyne of Rio Tinto. It involves ore loss and contact
Figure 7 shows the transformed distributions for four dilution. A good description is contained in Stone
possible SMUs: (1986), who spent many years collecting data and
SMU15561066 m (typical of a 3000 tpd operation) developing prediction formulae. Figure 9 shows a sketch
SMU251061066 m (typical of a 10 000 tpd opera- of dilution and ore loss at contacts. If mining will tend to
tion) follow the dashed lines, both dilution and ore loss will
SMU3515615610 m (typical of a 40 000 tpd operation) occur. If mining overbreaks the lower contact (follows
SMU4520620615 m (typical of a 80 000 tpd opera- the dotted line), there will be no ore loss but dilution will
tion). occur.
It is important to note that the coefficients of To evaluate contact dilution and ore loss a simulator
variation (s/m in Figs. 6 and 7) for the SMUs lie was developed. This simulator has been successfully
between the coefficients of variation for the composites used to adjust blocks for the Greens Creek underground
and the resource model. Therefore, the resource model mine, located near Juneau, Alaska and the True
has a lower coefficient of variation than the coefficients North open pit mine, located near Fairbanks, Alaska.
of variation for all the candidate SMUs. Figures 10–12 show contact conventions, mixing zones
Figure 8 shows grade–tonnage curves. The tonnage and mining conventions. The mixing zone may be
for SMUs is usually less than the resource model thought of as a zone where the contact shows short

170 Mining Technology 2012 VOL 121 NO 3


Parker Reconciliation principles for the mining industry

7 Histograms for SMUs

8 Grade–tonnage curves

Mining Technology 2012 VOL 121 NO 3 171


Parker Reconciliation principles for the mining industry

Table 10 Tonnages, grades and F1 factors

Results at 1.5 g/t cut-off grade (assume 100 Mtonnes orezwaste deposit)

%tons Mt Grade (g/t) Contained Au (Moz)

Composites 26.2 26.2 4.062 3.421


SMU15561066 m 31.1 31.1 3.229 3.228
SMU251061066 m 31.5 31.5 3.167 3.207
SMU3515615610 m 33.0 33.0 2.959 3.139
SMU4520620615 m 34.7 34.7 2.754 3.072
Resource model 37.1 37.1 2.487 2.966

F1 factors Tonnage Grade (g/t) Contained Au (Moz)


Composites 0.706 1.633 1.153
SMU15561066 m 0.838 1.298 1.088
SMU251061066 m 0.849 1.273 1.081
SMU3515615610 m 0.889 1.190 1.058
SMU4520620615 m 0.935 1.107 1.036
Resource model 1.000 1.000 1.000

Table 11 Annual cash flows for selectivity scenarios (base case shown in bold)

Input variables

SMU1 SMU2 SMU3 SMU4 Resource Model

Mining costs/$/t 2.00 1?80 1.60 1.40 1.80


Treatment costs/$/t 25.00 25?00 25.00 25.00 25.00
GzA costs/$/t 3.50 3?50 3.50 3.50 3.50
Sustaining Capex/$/t 0.30 0?27 0.24 0.21 0.27
Metallurgical recovery/% 88 88 88 88 88
Gold price/$/oz 1250 1250 1,250 1,250 1,250
Annual pretax cash flow (millions $); Assumes 10 Mtpa orezwaste

SMU1 SMU2 SMU3 SMU4 Resource model

Ore mined/Mt 3.11 3?15 3.3 3.47 3.71


Ore grade/g/t 3.23 3?17 2.96 2.75 2.49
Contained MOz 0.323 0?321 0.314 0.307 0.297
Recov. MOz 0.284 0?282 0.276 0.270 0.261
Revenues/M $ 355.1 352?8 345.3 338.0 326.3
Mining/M $ 220.0 218?0 216.0 214.0 218.0
Treatment/M $ 277.8 278?8 282.5 286.8 292.8
GzA/M $ 235.0 235?0 235.0 235.0 235.0
Sustaining Capex/M $ 23.0 22?7 22.4 22.1 22.7
Cash flow/M $ 219.4 218?3 209.4 200.1 177.8

amplitude deviations from a plane or a zone of mixing In the hypothetical case, the focus is on the sensitivity
during blasting. of dilution and ore loss to the degree of mixing and
Ore loss and dilution are expressed in distance in mining face position. Held constant are the strike
metres perpendicular to block faces on contacts. For an and dip of the contact (0u and 45u) and the slope of
ore block: the mining face (70u). Table 12 shows the simulator
output.
Adjusted tonnes~tonnes in situ{face area|Densityore | Where the face position is more negative (compare
Ore loss (m)zface area| runs 3, 8, 13, 18 and 23) there is more dilution and less
ore loss. For a given face position (compare runs 11, 12,
Densitywaste |Dilution (m) 13, 14 and 15) there is more ore loss and dilution as the
Adjusted grade~½tonnes in situ|grade in situ{ half width (TOLMIX) of the mixing zone is increased
from 1 to 5 m.
face area|Densityore |Ore loss (m)|
Table 13 shows production results. For simplicity the
grade in situzface area|Densitywaste | short term model is assumed to follow a tabular bed of
Dilution (m)|dilution grade=adj: tonnes iron ore in the field. Table 14 shows the simulator cases
and the resultant F2 factors. Run 15 (mixing is 5, mining
Similar formulae apply to transfer of ore loss into face position is 0) has F2 factors that match the
adjacent waste blocks. production results (Table 13).

172 Mining Technology 2012 VOL 121 NO 3


Parker Reconciliation principles for the mining industry

9 Ore loss and contact dilution (after Stone, 1986)

Table 14 shows the calculated NPVs, which take into Conclusions


account not only run of mine grade but an adjusted mine
life (10 Mtpa run of mine assumed) related to the net In most mines the most critical factors in poor
amount of ore loss/dilution gain. These have been recast reconciliation relate to lack of orebody knowledge or
in Table 15, with the optimal case for each mixing zone failure to mine to the defined mine plan. In the first case,
shown in italic. If the mixing remains at 5 m, the highest the plan is based on an incorrect interpretation of the
NPV will come by moving the face position west to 25. orebody in terms of orebody position, local tonnage and
The contact will be overbroken, but ore loss will be grade, or degree of selectivity implicit in the long term
minimised. If mixing can be reduced, the degree of Table 13 Production statistics
overbreak required can be reduced.
Mtonnes %Fe Mt cont. Fe
The observation made by Schofield (2001) is again
made clear. The reconciliation should be taken as a basis Short term model 10.00 40.00 4.00
for improvement. In a real life situation, the grade at the depletion
contact would need to be considered, with the amount of Received at mill 9.75 38.00 3.71
overbreak being dependant on the ore grade. F2 factor 0.975 0.950 0.926

Table 12 Simulator Output

Run Mixing Mine face position Dilution Width (m) Ore Loss Width (m)

1 1 25 4.94 0.00
2 2 25 5.00 0.03
3 3 25 5.37 0.27
4 4 25 5.62 0.58
5 5 25 6.10 1.06
6 1 22.5 2.58 0.17
7 2 22.5 2.89 0.46
8 3 22.5 3.53 0.96
9 4 22.5 4.13 1.62
10 5 22.5 4.74 2.24
11 1 0 0.91 0.97
12 2 0 1.30 1.33
13 3 0 2.23 2.13
14 4 0 2.83 2.80
15 5 0 3.52 3.49
16 1 2.5 0.17 2.76
17 2 2.5 0.40 2.96
18 3 2.5 1.01 3.44
19 4 2.5 1.55 4.04
20 5 2.5 2.25 4.76
21 1 5 0.00 5.06
22 2 5 0.03 5.06
23 3 5 0.27 5.17
24 4 5 0.56 5.52
25 5 5 1.11 6.08

Mining Technology 2012 VOL 121 NO 3 173


174
Parker

Mining Technology
2012
Table 14 Calculation of F2 factors for simulation runs*

From simulator Run of mine F2 factors NPV

VOL
Short term
Mine face Ore grade/ Density Diluted Density reserve Horiz. Dilution Ore loss Width/ Grade Density/

121
Run Mixing position %Fe of ore grade/%Fe of waste depl. Mt width ore/m width width m %Fe t/m3 Mtonnes Tonnes Grade Metal $M
Reconciliation principles for the mining industry

NO
1 1 25 40 3.6 0.1 2.6 10 40 4.94 0.00 44.94 37.50 3.490 10.89 1.089 0.939 1.022 1235

3
2 2 25 40 3.6 0.1 2.6 10 40 5.00 0.03 44.97 37.51 3.489 10.90 1.090 0.938 1.022 1233
3 3 25 40 3.6 0.1 2.6 10 40 5.37 0.27 45.10 37.33 3.481 10.90 1.090 0.933 1.017 1224
4 4 25 40 3.6 0.1 2.6 10 40 5.62 0.58 45.04 37.20 3.475 10.87 1.087 0.930 1.011 1215
5 5 25 40 3.6 0.1 2.6 10 40 6.10 1.06 45.04 36.95 3.465 10.84 1.084 0.924 1.001 1200
6 1 22.5 40 3.6 0.1 2.6 10 40 2.58 0.17 42.41 38.66 3.539 10.42 1.042 0.966 1.007 1256
7 2 22.5 40 3.6 0.1 2.6 10 40 2.89 0.46 42.43 38.50 3.532 10.41 1.041 0.962 1.002 1246
8 3 22.5 40 3.6 0.1 2.6 10 40 3.53 0.96 42.57 38.16 3.517 10.40 1.040 0.954 0.992 1229
9 4 22.5 40 3.6 0.1 2.6 10 40 4.13 1.62 42.51 37.84 3.503 10.34 1.034 0.946 0.978 1208
10 5 22.5 40 3.6 0.1 2.6 10 40 4.74 2.24 42.50 37.51 3.488 10.30 1.030 0.938 0.965 1188
11 1 0 40 3.6 0.1 2.6 10 40 0.91 0.97 39.94 39.50 3.577 9.92 0.992 0.988 0.980 1257
12 2 0 40 3.6 0.1 2.6 10 40 1.30 1.33 39.97 39.29 3.567 9.90 0.990 0.982 0.973 1245
13 3 0 40 3.6 0.1 2.6 10 40 2.23 2.13 40.10 38.78 3.544 9.87 0.987 0.969 0.957 1217
14 4 0 40 3.6 0.1 2.6 10 40 2.83 2.80 40.03 38.44 3.529 9.81 0.981 0.961 0.943 1196
15 5 0 40 3.6 0.1 2.6 10 40 3.52 3.49 40.03 38.05 3.512 9.76 0.976 0.951 0.929 1173
16 1 2.5 40 3.6 0.1 2.6 10 40 0.17 2.76 37.41 39.90 3.595 9.34 0.934 0.998 0.932 1225
17 2 2.5 40 3.6 0.1 2.6 10 40 0.40 2.96 37.44 39.77 3.589 9.33 0.933 0.994 0.928 1218
18 3 2.5 40 3.6 0.1 2.6 10 40 1.01 3.44 37.57 39.41 3.573 9.32 0.932 0.985 0.919 1201
19 4 2.5 40 3.6 0.1 2.6 10 40 1.55 4.04 37.51 39.09 3.559 9.27 0.927 0.977 0.906 1181
20 5 2.5 40 3.6 0.1 2.6 10 40 2.25 4.76 37.49 38.68 3.540 9.22 0.922 0.967 0.891 1157
21 1 5 40 3.6 0.1 2.6 10 40 0.00 5.06 34.94 40.00 3.600 8.74 0.874 1.000 0.874 1174
22 2 5 40 3.6 0.1 2.6 10 40 0.03 5.06 34.97 39.98 3.599 8.74 0.874 1.000 0.874 1174
23 3 5 40 3.6 0.1 2.6 10 40 0.27 5.17 35.10 39.83 3.592 8.76 0.876 0.996 0.872 1169
24 4 5 40 3.6 0.1 2.6 10 40 0.56 5.52 35.04 39.65 3.584 8.72 0.872 0.991 0.865 1157
25 5 5 40 3.6 0.1 2.6 10 40 1.11 6.08 35.03 39.31 3.568 8.68 0.868 0.983 0.853 1138
*Shown in bold is the case that most corresponds to production staitistics (see Table 13).
Parker Reconciliation principles for the mining industry

10 Contact conventions

model. In nearly all cases [Jerritt Canyon, Hayden Hill


and Goldstrike (USA); Kelian, Indonesia; Escondida,
Chile; Antamina, Peru; Diavik, Canada, and Palabora,
RSA are examples), the solution is infill drilling, and

12 Mining conventions

wherever possible blocks should be informed so that


they can be considered as proved ore reserves well in
advance of mining. In the second case, given that the
plan has been optimised, failure to mine to it will in the
end have the consequence of destroying value.
Reconciliation data can also serve as a basis for
improvement of the mining process. Examples are
Greens Creek Alaska (improved face sampling),

Table 15 Net present value presented as case matrix*

Mining face position

Mixing 25 22.5 0 2.5 5

1 1235 1256 1257 1225 1174


2 1233 1246 1245 1218 1174
3 1224 1229 1217 1201 1169
4 1215 1208 1196 1181 1157
5 1200 1188 1173 1157 1138
*Shown in bold is the case that most responds to production
statistics, shown in italic is the optimal case for a given mixing
zone half width (TOLMIX). The overall optimum case has mixing
11 Mixing zones of 1, mining face position of 0.

Mining Technology 2012 VOL 121 NO 3 175


Parker Reconciliation principles for the mining industry

Nchanga, Zambia (reduced unplanned dilution in both Morrison, R. D. 2008. An introduction to metal balancing and
reconciliation, 618, Brisbane, Qld, Julius Kruttschnitt Mineral
open pit and underground operations) and Stillwater,
Research Centre.
Montana (reduced mixing of ore and waste in remuck Parker, H. M. 1992a. The assessment of recoverable reserves for the
bays). In many cases these improvements come at very Minifie deposit using conditional simulation, executive summary,
little additional cost. prepared for Kennecott Explorations (Australia Ltd), Mineral
Resources Development Inc.
Parker, H. M. 1992b. Jerritt Canyon district reserve methodology,
Acknowledgement prepared for Independence Mining Company, Mineral Resources
Development, Inc.
The author would like to thank Compania Minera Parker, H. M. 2006. Resource and reserve reconciliation procedures for
Antamina, Rio Tinto and BHP Billiton for permission open-pit mines, AMEC report, 39.
to use some of the data and cases presented herein. Schofield, N. A. 2001. The myth of mine reconciliation, in Mineral
resource and ore reserve estimation – the AusIMM guide to good
practice, (ed. A. C. Edwards), 601–610, Melbourne, Vic.,
References AusIMM.
Stone, J. G. 1986. Contact dilution in ore reserve estimation, in Applied
Journel, A. G. and Huijbregts, C. J. 1978. Mining geostatistics, 600, mining geology: ore reserve estimation, (ed. D. E. Ranta), 55–166,
London, Academic Press. Littleton, SME.
Lane, K. F. 1988. The economic definition of ore, 149, London, Mining Yennamani, A. L., Aquirre, S. and Mousset-Jones, P. 2011. Blast-
Journal Books. induced rock movement measurement for grade control, Min.
La Rosa, D. and Thornton D. 2011. Blast movement modeling and Eng., February, 34–39.
measurement, Proc. 35th APCOM Symp., Melbourne, Vic., Zhang, S. 1998. Multimetal recoverable reserves estimation and its
Australia, September, AusIMM, 297–310. impact on the Cove ultimate pit design, Min. Eng., July, 73–79.

176 Mining Technology 2012 VOL 121 NO 3

You might also like