STOPE DESIGN CONSIDERATIONS USING ROCK MASS CLASSIFICATION TOOLS

AT THE XSTRATA ZINC GEORGE FISHER MINE

By Geoff W. Capes,1 Jeremy P. Doolan,2 and Leigh B. Neindorf3

ABSTRACT
MOUNT ISA

Darwin

Cairns
Townsville
NT
QLD

eal

HILTON
MINE

Brisbane

WA

hway
y Hig
Barkl

The development and use of rock mass classification

tools have been a key component for improving stope
design over the past 4 years at the Xstrata Zinc George
Fisher Mine in northern Queensland, Australia. In 2003,
stope extraction data from 3 years of open-stope mining
provided an excellent situation to review the assumptions
in the feasibility study. Extracted stope profile information,
drillhole geotechnical data, underground observations, and
oral and written communication were used to develop a
thorough stope reconciliation performance database. Without collecting the back analysis data and presenting the
data in a usable format, engineers are left to debate opinion
instead of engineered judgment. This can lead to biased
and uninformed design parameter choices with the potential to repeat poor design. This paper demonstrates some
effective, practical examples of empirical data collection
where rock mass classifications tools were developed and
used to create improved confidence in predicting stope
stability and failure profiles. The work contributed to
design changes that resulted in a reduction in stope
hangingwall dilution and an increase in head grade while
continuing to ramp up production from 20032005.

Cam
o ow

GEORGE
FISHER

SA
NSW
Perth

Airport

Mount
Isa

Sydney
Adelaide

AUSTRALIA

VIC

Canberra

Melbourne
TAS

ISA
MINE

Hobart
Boulia

CBD
Cloncurry

Power
Station

Figure 1.Location of George Fisher Mine.

collection of historic stope performance, rock mass classification, and underground observations of D ore body. The
evolution of the work has created a better understanding of
stope HW stability and created a change in mining
methodology. The work contributed to a reduction in
annual stope dilution from 14.4% (2003) to 6.3% (2005),
an increase in zinc grade from 7.4% to 8.7%, and an
increase in production from 2.1 to 2.6 Mt [Capes et al.
2006].
BACKGROUND

INTRODUCTION

Geology

This paper describes the development and use of rock

mass classification tools in stope design over the past
4 years at the George Fisher (GF) North Mine. The deposit
is located 22 km north along strike from the Mount Isa
Lead Mine. A joint stope dilution study between the
Xstrata Zinc George Fisher Mine (Figure 1) and the
University of Saskatchewan, Canada, began in October
2003. The objective was to create a thorough understanding of hangingwall (HW) overbreak using a sound
geological engineering approach that focused on data

The George Fisher Mine contains two similar deposits

located approximately 2 km apart with an estimated total
resource of 127 Mt at 100 g/t silver, 5.3% lead, and 9.2%
zinc (June 2005). The GF North deposit makes up approximately two-thirds of the current production, with the
remaining ore coming from GF South (previously known
as Hilton). The GF North deposit, the focus of this study,
is composed of a series of stratiform ore bodies striking
near north-south and dipping west between 30 to 70. The
sphalerite-pyrite-galena-pyrrhotite type mineralized areas
are separated by various thicknesses of bedded shales and
siltstones [Forrestal 1990].
D ore body is currently the key area of extraction,
grading at approximately 5% lead, 9% zinc, and 100 g/t
silver with the mineralization being unequally distributed
through the ore body, including a high-grade massive
sulfide layer sometimes defining the ore/HW contact.
Typically, the D ore body HW rock is composed of a
siltstone marker rib followed by a package of pyritic shales

1
Ph.D. candidate, Department of Civil and Geological Engineering, University of Saskatchewan, Saskatoon, Canada, and
senior planning engineer, Xstrata Zinc, George Fisher Mine,
Mount Isa, Queensland, Australia.
2
Rock mechanics engineer, Xstrata Zinc, George Fisher Mine,
Mount Isa, Queensland, Australia.
3
Project manager, Xstrata Zinc, Mount Isa, Queensland,
Australia.

37

Haydon 2006]. The two main types of faults are crosscutting north-east trending faults and bedding parallel
faults having their own individual characteristics and zone
of influence.

(5%20% fine-grained pyrite) and a series of very fissile

black shales (Figure 2). This is topped off with a variably
thick mineralized lead-zinc lens. A section of siltstone
separates this lens from the massive pyrite marker and
massive sulfides, which define the start of C ore body. The
thickness and rock mass quality of each package between
D ore body and C ore body vary quite significantly along
strike and dip as a result of past brittle and ductile
deformation from faulting and folding events. A large
number of geological structures affect the George Fisher
deposit with varying degrees of offset, ore body rotation,
drag folding, and metal redistribution [Grenfell and

Stope Layout and Design

The mining method at GF consists mainly of
transverse open-stope mining where the ore bodies are
greater than 10 m in thickness. Transverse stopes are
mined from a footwall drive access either on a 30- or 60-m
sublevel spacing giving consideration to the local rock
mass quality and stope shape relationship. Primary stopes
are mined until uneconomic conditions exist and then are
filled with either a cement aggregate fill or a paste fill
material. A secondary stope is typically mined once the
adjacent primaries have been mined to one 30-m sublevel
higher then the secondary stope and the fill has been
allowed to sit for 28 days and gain sufficient strength to act
as a sidewall during extraction (Figure 3). The original
1520 design (17.5-m crosscut centers), primary and
secondary stope strike lengths, respectively, was changed
in 2003 to a 1020 design (15-m crosscut centers). This
eventually became a floating design based on local rock
quality where primary stopes could be up to 15 m on strike
considering the cost of cemented backfill, local ore body
width, HW stability, and effects on the mining cycle. In
addition, the drill design of stopes within the George
Fisher Mine underwent a change in 2003 when D ore body
HW drives were eliminated and transverse stope crosscuts
were used as the only access for D ore body extraction.
One of the main motives behind this design change was

Faults
10 Level

11 Level
12 Level
60 m
13 Level
West

Figure 2.Typical cross-section of George Fisher ore

bodies.

Faults

C-orebody

HW Drive

Stope B
West

D-orebody

Transverse Cross cut

Stope Crosscuts
Stope A

Transvers Cross cut

FW Drive

HW and FW

Figure 3.George Fisher mining layout (plan and cross-section views).

38

the existence of problems with the stope hangingwalls and

adjacent hangingwall drive failures as a response to stoping. The elimination of HW drives created the need for
additional diamond drilling of the ore body to delineate the
stope extraction wireframes. This addition of drilling and
subsequent geotechnical logging has provided substantial
relevant information for stope design and reconciliation
purposes.
Numerical modeling provided an adequate set of rules
for mine-wide sequencing [Beck 2003], which, when
adhered to, lead to a sustainable mining method. The suggested guidelines included minimizing overbreak in
primaries to avoid connection of voids when mining
secondaries, tight-filling stopes, eliminating triple-lift
pendant pillars, decreasing stope cycle time, and creating a
better sequence to minimize stress effects on extraction.
However, even with the mine-wide changes and minewide sequencing rules, a set sublevel spacing and crosscut
width in poor-quality HW rock mass continued to result in
unpredicted major individual stope HW failures and subsequent production inefficiencies. Thus, a thorough understanding of individual stope performance was required in
order to achieve further positive design change.

DEVELOPMENT AND USE OF

ROCK MASS CLASSIFICATION TOOLS
AND EMPIRICAL METHODS FOR DESIGN
Overview
During late 2003, stope extraction data from 3 years of
mining provided an excellent situation to review the
assumptions in the feasibility study. Stope profile information based on data acquired using the cavity monitoring
system [Miller et al. 1992], drillhole (BQ size) geotechnical data, underground observations, and oral and
written communication was used to develop a thorough
stope reconciliation performance database. On-site
research, coupled with subsequent design trial and implementation, resulted in the development of a model that
showed a relationship between rock mass quality, span,
and resultant extraction profile. The goal of this model was
to use it as a template and continually update the knowledge with underground observations and reconciliations of
case histories. The model represented significant time, discussion, and research, but with more data, the model
would evolve and lead to continued improvements in the
understanding of HW behavior and design of stopes.

Figure 4.RQD long section contoured.

39

Phase 1: RQD Long Section

the required hydraulic radius (H.R.), modified stability

number (N ) ranges.
However, further work was required to examine why
some cases did not agree with the design approach based
on the average rock mass conditions. When Category 3
rocks (RQD < 10%20%, N < 1) were analyzed, the failure prediction did not seem to correlate with the ELOS
predictions based on a 5-m average of HW conditions.
Two such examples included having a 6-m ELOS in a
primary stope where <0.5 m was expected (N = 0.5, H.R.
= 3) and an 13-m ELOS where approximately 4 m was
expected (N = 0.8, H.R. = 7). For category 1 (RQD
> 40%, N > 6) and category 2 (RQD 20%40%, N = 14)
rocks, the majority of cases of ELOS prediction were
within acceptable error for stope prediction requirements,
but an improved model was desired to examine the cases
that did not fit.

The first step of the study was to create a long section

of the available geotechnical data. The comprehensiveness
and density of geotechnical data depend on the year in
which the drillhole was logged. Most holes after 1999 have
RQD data available (BQ: 36.540.7 mm) as recommended
by Hadjigeorgiou [1999]. Rock Quality Designation (RQD)
[Deere 1964] was originally developed for drill core 54
mm, although the BQ RQD is used to compare site data.
The RQD data were averaged for the first 5 and 10 m of
rock into the hangingwall and placed on a long section
showing significant variation along strike and dip. The 5-m
average RQD long section (Figure 4) has evolved from a
scrappy, hand-contoured, coffee-stained desk map to a
sharable plan able to be accessed by the technical services
team on the mine design software. The original section
was hand-contoured while the up-to-date section was contoured by triangulating the RQD values on the mine software as topographic surfaces between points and joining
equal elevation points as contours. There is a substantial
increase in RQD data density above the 2,800 elevation
due to the removal of the HW drives and requirement for
diamond drilling and subsequent geotechnical logging.

Category1
RQD > 40%
(for at least 1-2m)
Jn < 6-9
Jr = 1
Ja = 1
Q = 4.4

Phase 2: Empirical Methods

A =1, B=0.3, C= 4-7

The next step was to gather additional averaged data

for use in empirical design methods, such as the modified
stability graph method [Potvin 1988] and dilution graph
method [Clark and Pakalnis 1997], which require a single
estimate of rock mass properties for the surface being
analyzed. This involved acquiring geotechnical parameters
such as joint condition, joint alteration, and joint roughness
used in the Q-system [Barton et al. 1974]. Based on many
underground observations and stope reconciliations, the
rock in the HW area was assigned three broad categories
(Figure 5). Each stope was assigned values within a category and plotted with actual ELOS values using the
dilution graph format. ELOS (equivalent linear
overbreak/slough) is defined as the volume of the HW
overbreak divided by the stope HW surface area [Clark
and Pakalnis 1997]. ELOS is a useful tool for mine planning because it provides the ability to quantify a diluted
stope shape to enhance mine scheduling purposes. The
dilution graph was modified with new ELOS lines calibrated to data collected for stoping in poor-quality rock
masses where minimal data have been presented except for
Capes et al. [2005] and Brady et al. [2005] (Figure 6).
The calibrated curves on the modified dilution graph
(Figure 6, right) have been used to predict HW overbreak
in both GF North and South operations for the last few
years. The calibrated curves, which need to be statistically
verified as a final part of the study, are an effective tool for
local stope design as they contain a large amount of data in

N > 6
Category 2
RQD= 20-40%
(5m average)
Jn= 12
Jr = 1
Ja = 2
Q = 0.8-1.6
A =1, B=0.3, C= 4-7

N = 1 - 4

Category 3
RQD<10-20%
(5m average)
Jn= 12-15
Jr = 1-2
Ja = 2-4
Q = 0.3-0.8
A =1, B=0.3, C= 4-7

N < 1
Figure 5.Examples of HW rock categories at George
Fisher Mine.

40

Dilution Graph

Modified Dilution Graph

George Fisher Rock Category

Figure 6.Dilution graph (from Brady et al. [2005]) and modified dilution graph with additional data from poor-quality
rock masses (after Capes et al. [2005]).

stability relationship between the composite beam

thickness (meters of BQ core >40% RQD) and span
(Figure 8). Figures 912 demonstrate four case histories of
the model incorporating the three rock categories.
Recognizing the changing rock mass condition with

Phase 3: Rock Category and Failure Profile

Relationships
The next step was to investigate the additional factors
that have been documented to relate to HW overbreak
from a comprehensive literature review and studies conducted at the George Fisher Mine. These additional factors, including faulting, stress, blasting, undercutting, and
time, were identified to examine how or if they played a
role in erroneous ELOS prediction. During this next phase
of the study, interesting relationships were observed
between failure shape, rock quality, and span when the
RQD drillhole data were overlain as logged intervals on
HW failure profiles in the mine design software. Thus, the
methodology was changed to first investigate these novel
relationships before examining the additional factors.
RQD cross-section plots were created for individual
stopes and provided valuable data for understanding the
variability of rock quality near the HW/ore boundary
(Figure 7). The RQD plots coupled with underground
observations created the idea of different failure profiles
(Figure 7) for the different rock mass categories shown in
Figure 5. Areas where the RQD < 10%20% typically
failed into the next rock mass category without arching.
Areas with RQD 20%40% would arch to become stable
within span constraints or would change profile when a
different rock category was intersected. Stability existed in
other stopes where the stope failed to a composite beam or
plate of rock in which the RQD > 40% for at least 1 m.
These relatively thin zones of higher RQD rock define a
more stable domain, and the position of this domain was
found to define the extent of failure, within span
constraints. These stable domains were not always a
consistent lithology, but could often be defined as pyritic
shales, siltstone beds, or a narrow mineralized area. Data
were collected to create a design tool demonstrating the

Representative RQD Plot

70
Orebody

Hangingwall

60

50

RQD%

40

30
20

10
0
-10

-5

Distance from HW(m)

10

Failure
Profile
N
15 m

Plan View

HW

Arching

Stable
Beam/Plate

Figure 7.RQD plot and representative failure

shape profiles based on rock category and span.

41

Span (m)
sublevel height = 30 m, dip =40-70

25

West

20

Section View

20-40% < 20%

FAILED

15

10

HW

STABLE

FW

5
Beam stable under span
Beam failed under span
0
0

Beam thickness (m) where beam RQD > 40%

West

Figure 8.Composite beam thickness versus span stability graph.

>40%

20%

D-FW

Plan View
D-HW

Section View

D-FW

Drillhole

<20% 20-40%

D-HW

West
RQD Plot - 742D 10L-10C (7440N Upper)
70

Plan View

20%

ORE HW

>40%

60
50
40

RQD %

Drillhole
D-FW

30

D-HW

20
10
0

West

-10

-5

10

Distance from HW (m)

Figure 10.742D 10L10C. Primary stope where 6- to

8-m depth of HW failure was predicted. Failure depth was
810 m. Stope failed without arching through first 5 m of
category 3 rock, then arched off to a stable profile through
category 2 rock (September 2006).

RQD Plot 716D 12C-11L (7163N Central)

70
ORE HW
60
50

distance from the hangingwall contact enabled an

improved and accurate method of estimating HW dilution.
RQD was noted as the most significant contributor as a
geotechnical input into stope design and is the most
available and most easily obtained data. Stope performance prediction now includes a cumulative overbreak prediction where the extraction profile is estimated through
the different domains to create an expected stope shape for
mine planning. Further work is being conducted to look at
correlations between arching failure angle, span, and RQD.
This may be difficult to quantify due to the lack of available data to estimate bedding perpendicular joint spacing
for individual stopes.

RQD%

40
30
20
10
0
-10

-5

10

Distance from HW Contact(m)

Figure 9.716D 12C11L. Secondary stope where a

major failure was expected using predicted stress conditions. Local rock mass conditions limited depth of failure
to 23 m as stope failed through category 3 rock to the
category 1 composite beam (February 2004).

distance from the hangwall contact enabled an improved

42

11-m HW span

17-m HW span

NOTE: Dot represents similar position under each span. Span was increased in the upper left of photo.

Plan View

West
D-HW

D-FW

Second Beam

>40% <20% 20-40%

First Beam

Drillhole

RQD Plot 742D 12C-11L(7440N upper)

First Beam

70

Second
Beam

60
50
Orebody Hangingwall

RQD%

40
30
20
10
0
-10

-5

10

Distance from HW(m)

Figure 11.742D 12C11L. First composite beam (category 1) was stable under 11-m span, but failed under a 17-m span,
arching through the category 2 rock to the second composite beam (May 2004).

43

Category 1 Rock

Category 3 Rock

Figure 12.Looking west through the open stope. Photo shows failure through category 3 to category 1 rocks.

Phase 4: Additional Factors

available to show the effect of using low- or high-density

explosive near a category 1 rock mass. This is shown as
the explosive-sensitive zone on the composite beam versus
span stability plot (Figure 8).
The performance of primary and secondary stopes was
compared to examine the effect of the different stress conditions under which the stopes were mined. There were no
consistently observed trends to indicate that stress levels
influenced stope HW behavior on a mine-wide scale based
on a comparison of depth of HW failure in primary stopes
and secondary stopes in the same area of the mine. The
comparison of depth of failures averaged 4.2 m for secondary stopes (58 samples) and 4.3 m for primary stopes
(75 samples). However, secondary stopes have most often
either performed in a stable manner or had >5-m depth
failure, with significantly fewer cases in the 0- to 5-m
depth of failure category (Figure 14). The sizes of primary
and secondary stopes have varied through the life of the
mine, providing a good spread of data for comparison.
Stope HW performance can be seen as a function of individual stope design choices, where the stope depth of
failure is plotted on a long section with rock quality
(Figure 13). Secondary stopes extracted with adequate
sequencing have performed well under larger HW spans
than primary stopes where the HW RQD in the secondary
stope has been better than that of the primary. On the other
hand, there have been additional HW and back failures in
secondary stopes where mining has been conducted out of
sequence (triple-lift pendants), where cemented fill has
failed in primaries, and some areas where primary stopes
in the lift above were not filled for up to 8 months,
resulting in additional fall-off and a subsequently worse

With the improved model for predicting HW

performance developed, the effect of faulting, blasting,
stress, undercutting, and time were then investigated to
examine how much of a role they played in stope HW
stability.
The effect of faults at the mine was debated, as some
stopes with faults passing through had major overbreak,
while others remained stable under similar stoping conditions. Some trends between faulting and lower RQD values
were evident, but were not consistent. The RQD long section showed that RQD and stope stability could not consistently be related to the location of faulting, although
poor rock quality can be associated with some of the faults
(Figure 5). For example, there has been failure next to the
S73 fault on the majority of levels where the stopes have
been mined to the north of the fault (Figure 13). This correlation cannot be seen with all the north-east trending
faults and agrees with the geological descriptions that the
major faults show variable dextral offset, displacement and
features [Grenfell and Haydon 2006].
Blasting has not been analyzed to the same degree as
the other factors. This is due to the consideration of a 25-mwide ore body, the failures are occurring at a depth much
greater than the influence of ANFO higher-density explosive (approximately 0-2 m back break based on local opinion) or ISANOL lower-density explosive (approximately
0- to 0.5-m back break) in the HW ring of an 89-mm-diam
production hole, and that the majority of HW rings have
been charged with low-density explosive. The study has
only touched on a few areas where data have been

44

Figure 13.Long section with faults, stopes, and depth of failure (plotted at stope centerlines).

failure and caved 20 m above the designed stope level

when it was predicted to have 23 m of HW failure based
on the cumulative overbreak model. The primary stopes
had all been mined at a 15-m strike span to access the
tonnes sooner and were left open for up to 254 days, during which time sidewall failures occurred off of the secondary stope pillars, resulting in a stope <15 m wide. It is
important to capture such information to verify the limits
to which the design model can be applied and to reiterate
the potential downstream mining-influenced effects where
sequencing and backfill conflict with meeting production
targets.
Current work is being completed to determine the
effects of producing stopes adjacent to voids that have
remained open for extended periods of time. The mining
cycle, from firing to filling, is being analyzed with respect
to stope behavior prediction and performance, with emphasis on time-related failure of the stopes. Analysis to date
has shown that relaxation/failure of the secondary pillar
seems to be reduced by the timely filling of the primary
stopes on the same level as well as the next lift above.

30

Primary
Secondary

Number of Cases

25

20

15

10

0
< 0 m overbreak

0-5 m Depth of failure

> 5 m depth of failure

Depth of Failure Category

Figure 14.Comparison of failures in primary and

secondary stopes.

stress condition for mining the secondaries. During 2006,

significant delays in cement filling and/or lack of tight filling of previous lifts of the primaries were experienced.
Three of the four secondary stopes (719D, 723D, 730D
10L10C) experienced large back failure/caving above the
designed stope level. One particular stope had 10 m of HW

45

overbreak. This was achieved by creating a model for

design based on historical performance, understanding
why it worked and why it did not work, appreciating it for
the insights it provided, and continually reanalyzing as
more information became available. The study has
reemphasized the belief in understanding relevant factors
for individual stope design. The averaging approach for
estimating rock mass properties for design methods does
not always provide the best answer based on available
information. Many considerations must be given to design
individual stopes. Generalizing all stope designs into one
category can lead to poor design. Individual fault characteristics play a role in stope stability. There is no discernible difference in average HW depth of failure between
primary and secondary stopes, although mining out of
sequence, triple-lift pendants, and poor cement fill quality
or lack of tight fill and/or delayed fill in primary stopes
may lead to greater HW failure and create subsequent
stope back failure. The model created the ability to
develop innovative empirical stope design tools to control
and reduce dilution in conjunction with following a set of
mine-wide extraction rules developed from numerical
modeling. The development and use of the model resulted
in significant research and development benefits to the
company. The confidence in prediction created individual
stope design changes that contributed to a reduction in
dilution and an increase in head grade while continuing to
ramp up production from 20032005.

Being able to recognize potential poor stope performance

at an early stage in the design process will allow for more
efficient mine planning and extraction, with manageable
dilution. So far, telltale signs of poor-performing stopes
have been mining difficulties when establishing secondary
crosscuts, such as arching failure of the crosscut profiles,
and difficulty installing ground support and reinforcement.
HW and/or back overbreak from the previous lift must also
be considered to determine stope HW design choices for
the next lift. In conjunction with the model, the timerelated failure study, together with observational data,
should give the planning team more insight into predicting
future stope performance and managing stope stability.
Summary: Results of Trials and Methodology

16

8.8

14

8.6

12

8.4

10

8.2

7.8

2.4 Mt

2.6 Mt

2.1 Mt

2003

2004

2005

0
2002

ACKNOWLEDGMENTS
The authors would like to thank Xstrata Zinc and the
University of Saskatchewan for their support, as well as
the various technical services and operations personnel and
consultants who have shared their thoughts during their
time at George Fisher Mine Operations.
REFERENCES

Zinc Grade%

Dilution%

Management and operational support to implement

changes to individual stope designs was very positive, and
many ideas were discussed to improve mine performance
based on the understanding of HW behavior. Based on the
cumulative overbreak model where the rock mass conditions meet the specific criteria, these design choices
include the verticalization of HWs, use of cabled and noncabled ore chocks for HW stability, use of ore skins where
mining consequences from HW failure are high, and an
effective rock mass management strategy [Capes et al.
2005]. The development and use of the model resulted in
significant research and development benefits to the company. Figure 15 shows the contribution of the stoping
methodology change as a reduction in annual stope dilution from 14.4% (2003) to 6.3% (2005), an increase in zinc
grade from 7.4% to 8.7%, and an increase in production
from 2.1 to 2.6 Mt [Capes et al. 2006].

Beck DA [2003]. Stress influences on rockmass damage and stope stability at GF mine-calibrated model criteria. Consultant report for Xstrata Zinc, George Fisher
mine.
Barton N, Lien R, Lunde J [1974]. Engineering classification of rock masses for the design of tunnel support.
Rock Mech 6(4):189236.
Brady TM, Pakalnis R, Clark L [2005]. Design in
weak rock masses: Nevada underground mining operations. SME preprint 0543. Littleton, CO: Society for Mining, Metallurgy, and Exploration, Inc.
Capes GW, Grant DR, Neindorf LB [2006]. George
Fisher: sustainable mine design. Paper presented at the
Second International Seminar on Strategic Versus Tactical
Approaches in Mining (Perth, Australia).
Capes GW, Milne DM, Grant DR [2005]. Stope
hangingwall design approaches at the Xstrata Zinc, George

7.6
Dilution%

7.4
Zinc Grade

7.2
2006

Year

Figure 15.George Fisher North stope dilution percent, total mine tonnes, and zinc grade versus time.

CONCLUSIONS
Dedicated on-site research using a sound geological
engineering approach of data collection, communication,
underground observations, analysis, and implementation of
ideas resulted in a thorough understanding of HW

46

Grenfell K, Haydon M [2006]. Challenges in modelling large complex orebodies at the George Fisher north
mine. In: Proceedings of the Sixth International Mining
Geology Conference (Darwin, Northern Territory, Australia, August 2123, 2006).
Hadjigeorgiou J [1999]. Memorandum of March 11,
1999, from John Hadjigeorgiou, Laval University, Quebec,
Canada: a discussion document on geotechnical rock mass
characterisation for Mount Isa.
Miller F, Jacobs D, Potvin Y [1992]. Laser measurement of open stope dilution. CIM Bull 85:Jul-Aug.
Potvin Y [1988]. Empirical open-stope design in
Canada [Dissertation]. Vancouver, British Columbia,
Canada: University of British Columbia.

Fisher mine, North Queensland, Australia. In: Chen G,

Huang S, Zhou W, Tinucci J, eds. Proceedings of the 40th
U.S. Rock Mechanics Symposium (Anchorage, AK, June
2529, 2005). Alexandria, VA: American Rock Mechanics
Association.
Clark L, Pakalnis R [1997]. An empirical design
approach for estimating unplanned dilution from open
stope hangingwalls and footwalls. Presented at the 99th
Canadian Institute of Mining Annual Conference (Vancouver, British Columbia, Canada).
Deere DU [1964]. Technical description of rock cores
for engineer-purposes. Rock Mech Eng Geol 1(1).
Forrestal PJ [1990]. Mount Isa and Hilton silver-leadzinc deposits. In: Hughes FE, ed. Geology of the mineral
deposits of Australia and Papua New Guinea. AusIMM
Monogr 14:927934.

47