University of Exeter, Camborne School of Mines

Determination of an Ultimate Pit Limit Utilising Fractal

Modelling to Optimise NPV

Submitted by Amir Bijan Yasrebi, to the University of Exeter

as a thesis for the degree of Doctor of
Philosophy in Earth Resources
November 2014

This thesis is available for library use on the understanding that it is copyright material
and that no quotation from the thesis may be published without proper
acknowledgment.

I certify that all material in this thesis which is not my own work has been identified and
cited and that no material has previously been submitted and approved for the award of
a degree by this or any other University.

1
Abstract

The speed and complexity of globalisation and reduction of natural resources on the one

hand, and interests of large multinational corporations on the other, necessitates proper

management of mineral resources and consumption. The need for scientific research and

application of new methodologies and approaches to maximise Net Present Value (NPV)

within mining operations is essential.

In some cases, drill core logging in the field may result in an inadequate level of

information and subsequent poor diagnosis of geological phenomenon which may

undermine the delineation or separation of mineralised zones. This is because the

interpretation of individual loggers is subjective. However, modelling based on logging

data is absolutely essential to determine the architecture of an orebody including ore

distribution and geomechanical features. For instance, ore grades, density and RQD

values are not included in conventional geological models whilst variations in a mineral

deposit are an obvious and salient feature. Given the problems mentioned above, a series

of new mathematical methods have been developed, based on fractal modelling, which

provide a more objective approach. These have been established and tested in a case

study of the Kahang Cu-Mo porphyry deposit, central Iran.

Recognition of different types of mineralised zone in an ore deposit is important for mine

planning. As a result, it is felt that the most important outcome of this thesis is the

development of an innovative approach to the delineation of major mineralised

(supergene and hypogene) zones from ‘barren’ host rock. This is based on subsurface

2
data and the utilisation of the Concentration-Volume (C-V) fractal model, proposed by

Afzal et al. (2011), to optimise a Cu-Mo block model for better determination of an ultimate

pit limit. Drawing on this, new approaches, referred to Density–Volume (D–V) and RQD-

Volume (RQD-V) fractal modelling, have been developed and used to delineate rock

characteristics in terms of density and RQD within the Kahang deposit (Yasrebi et al.,

2013b; Yasrebi et al., 2014). From the results of this modelling, the density and RQD

populations of rock types from the studied deposit showed a relationship between density

and rock quality based on RQD values, which can be used to predict final pit slope.

Finally, the study introduces a Present Value-Volume (PV-V) fractal model in order to

identify an accurate excavation orientation with respect to economic principals and ore

grades of all determined voxels within the obtained ultimate pit limit in order to achieve

an earlier pay-back period.

3
Acknowledgments

It would not have been possible to write this doctoral thesis without the help and support

of the kind people around me, to only some of whom it is possible to give particular

mention here.

I would like to express my special appreciation and thanks to my supervisors Dr. Andrew

Wetherelt, Dr. Patrick J Foster and Dr. Peyman Afzal for your tremendous support. I would

like to thank you for encouraging my research and for allowing me to grow as a research

scientist. Your advice on both research as well as on my academic behaviour have been

priceless. I would also like to acknowledge you for your brilliant comments and

suggestions, thanks to you.

I would like to be especially thankful to my first supervisor, Dr. Andrew Wetherelt, for the

patient guidance, encouragement and advice he has provided throughout my time as his

PhD student. I have been extremely lucky to have a supervisor like him who cared so

much about my work and who always responded to my questions and queries so

promptly.

I would like to thank the Institute of Materials, Minerals and Mining, the global network

IOM3, Cornish Institute of Engineers and Whittle Consulting (Business Optimisation for

the Mining Industry) not only for providing the funding which allowed me to undertake this

research, but also for giving me the opportunity to attend conferences and meet so many

interesting people.

4
I am grateful to the National Iranian Copper Industries Co. (NICICO) for their permission

to have access to the Kahang deposit dataset. Additionally, I would like to thank Dr.

Ardeshir. Saad Mohammadi (former CEO of NICICO), Mr. Reza Esfahani Pour (former

manager of Exploration and Development Engineering of NICICO), Mr. Ahmad

Moradalizadeh (current CEO of NICICO) and Mr. Babak Babaiee (current manager of

Exploration and Development Engineering of NICICO) for their support.

Above all, I would like to thank my parents who have given me their unequivocal support

throughout, as always, for which my mere expression of thanks likewise does not suffice.

Words cannot express how grateful I am to my parents for all of the sacrifices that they

have made.

At the end, the great appreciation to my blessed grandmother, her prayer for me was

what sustained me thus far.

GOD RICHLY BLESS YOU ALL

5
 Table of Contents

1 CHAPTER ONE. Introduction 26

1.1 Problem Description 27
1.2 Objectives of the Research 33
1.3 Application of Fractal/Multifractal Modelling with Reliance on Geochemical
Population 34
1.3.1 Introduction to Common Fractal Models 38
1.4 Methodology 39
1.4.1 Introduction to NPV Scheduler 40
1.4.2 Introduction to Lerch and Grossmann Algorithm with Reliance on Resource
Modelling 40
1.4.2.1 Introduction to Other Methods for Mine Planning 44
1.5 Specific Economic and Political Context of Mining in Iran 45
1.6 Organisation of the Thesis 48
2 CHAPTER TWO. Geology and Associated Mineralisation 50
2.1 Regional Geology 51
2.2 Geology of the Kahang Deposit 52
2.2.1 Mineralisation Characteristics of the Kahang Deposit 54
2.2.2 Hydrothermal Alteration 60
2.3 Dataset Particulars 68
2.4 Assay Quality Assurance and Quality Control 69
2.4.1 Comparison of Geochemical Data Variances via F-Distribution 72
2.4.2 Comparison of Geochemical Data Means via Paired T-Test 73
2.5 3D Geological Modelling of the Deposit 75
2.5.1 Lithological Model 75
2.5.2 3D Alteration Model 78
2.5.3 Ore-type Zonation Model 80
3 CHAPTER THREE. Concentration-Volume (C-V) Fractal Modelling for Separation of
Mineralised Zones 84
3.1 Introduction 85
3.2 Geometry of Natural Processes 86

6
 3.3 Statistical Characteristics 89
3.4 Block Modelling 96
3.4.1 Cell Declustering 102
3.5 Geostatistical Modelling 109
3.5.1 Inverse Distance Weighted Anisotropic Method (IDWAM) 110
3.5.2 Application of IDWAM 115
3.6 Validation Processes 133
3.7 Delineation of Mineralised Zones Using C-V Modelling 136
3.7.1 C-V Fractal Model 136
3.7.2 Application of C-V Fractal Modelling 143
3.8 Comparison and Correlation between Results of C-V Fractal and Geological
Modelling 146
3.8.1 Application of Logratio Matrix 155
3.9 Results 159
4 CHAPTER FOUR. Rock Mass Characterisation Utilising Fractal Modelling based on
Density and RQD Data 161
4.1 Introduction 162
4.2 Statistical Characteristics 166
4.3 Methodology 169
4.3.1 D-V Fractal Model 172
4.3.2 RQD-Volume (RQD-V) Fractal Model 177
4.4 Comparison between D–V and RQD Models 183
4.5 Results 185
5 CHAPTER FIVE. Mining Optimisation 187
5.1 Introduction 188
5.2 Methodology 192
5.3 Deposit Block Model via the C-V Fractal Model for Optimisation Study 193
5.4 Mine Topographical Features of Land Surface 194
5.5 Pit Geometrical Characteristics 195
5.6 Mine’s Annual Production 198
5.7 Ore Density 198

7
 5.8 Kahang’s Exploitation Percentage 201
5.9 Economic Principles 202
5.9.1 Prices and Expenses 202
5.9.2 Annual Discount Rate 204
5.9.3 Cut-off Grade 207
5.10 Determination of the Kahang Deposit Ultimate Pit Limit 208
5.10.1 Ultimate Pit Limit 208
5.10.2 Internal Pit Shells (Phases) 211
5.10.3 Nested Pits 214
5.10.4 Identification of an Optimal Extraction Sequence (OES) 214
5.11 Comparative Case Study 216
5.12 Determination of an Ultimate Pit Limit when Ignoring the Three Isolated
Boreholes 221
5.13 Results 224
6 CHAPTER SIX. Present Value-Volume (PV-V) Fractal Modelling for Mining Strategy
Selection 226
6.1 Introduction 227
6.2 Methodology 229
6.3 Statistical Characteristics 231
6.4 Application of PV-V Model 233
6.5 Application of NPV-CTO Model 245
6.6 Results 247
7 CHAPTER SEVEN. Conclusions and Recommendation for Future Work 249
References 259
Appendices 297

8
 List of Figures

Fig. 1. 1. An Example of the graph closure in the Lerch and Grossman algorithm

(Meagher et al., 2010) 43

Fig. 1. 2. Bulking the block model voxels together (Mart and Markey, 2013) 44

Fig. 2. 1. a) Geological map of the Kahang study area, scale: 1: 10,000 (Alavi, 1994;
Tabatabaei and Asadi Haroni, 2006), and b) structural map of Iran, showing the
Urumieh-Dokhtar volcanic belt (Alavi, 1994) 54

Fig. 2. 2. Eocene sub-volcanic rocks in the Kahang deposit (View towards SE) 55

Fig. 2. 3. a) Pyrite (Py), chalcopyrite (Ccp) and molybdenite (Mol), b) Copper secondary
sulphides and carbonate minerals from the oxidised zone (Azadi et al., 2014), c) Fe-
oxides, and d) reflected light photomicrograph showing bornite (Brn), chalcocite (Chl)
and chalcopyrite (Cpy) in the Kahang porphyry deposit 57

Fig. 2. 4. Photomicrographs of ore minerals in the Kahang deposit (Afshooni et al.,

2013): a) type II pyrite (Py II) associated with chalcopyrite, b) replacement of pyrite by
magnetite in type I pyrite (Py I), c) subhedral sphalerite, containing chalcopyrite
inclusions enclosed by pyrite, d) magnetite grains associated with Ti-mineral, pyrite and
chalcopyrite, e) ex-solution between chalcocite and chalcopyrite, f) hematite blades, g)
galena grains associated with chalcopyrite, h) covellite occurs as fracture-filling in pyrite,
i) malachite occurs as fracture-filling in micro-diorites, j) bornite together with
chalcopyrite, k) chalcocite, digenite and pyrite occur as veins, l) Backscattered electron
Image showing native gold (electrum) grains within late stage grey quartz 59

Fig. 2. 5. a) Alteration map of the Kahang deposit (Western, Central and Eastern parts;
Harati et al., 2013) and b) Conceptual model of Lowell and Gilbert (1970) 61

Fig. 2. 6. Photomicrographs of the potassic alteration zone of the Kahang deposit: a)

neo-formed biotite (Bt) and KF veinlets, b) secondary biotite (S-Bt) and quartz (Qtz), c)
secondary biotite–chlorite assemblage after igneous amphibole associated with the
potassic alteration (Afshooni et al., 2013), and d) fine-grained biotite as pseudomorphs

9
of amphibole phenocryst, and coarse-grained biotite cut by a quartz veinlet (Afshooni et
al., 2013) 63

Fig. 2. 7. Photomicrographs of phyllic alteration zone in the Kahang porphyry deposit: a)

presence of sericite (Ser) and quartz (Qtz), b) plagioclase (Plg) phenocrysts are
pervasively replaced by sericite and surrounded by quartz grains in the phyllic alteration
zone (Afshooni et al., 2013), and c) biotite (Biot) altered to chlorite (Chl) in the phyllic
alteration zone (Afshooni et al., 2013) 65

Fig. 2. 8. Photomicrographs of argillic alteration zone in the Kahang porphyry deposit: a)

argillic alteration with clay minerals, altered plagioclase (Plag) and opaque minerals
(Op), b) K-feldspar (Kf) phenocrysts partially replaced by clay minerals (fine grey
material) in argillic alteration zone (Afshooni et al., 2013), and c) Advanced argillic
alteration containing jarosite (Jar) overprinted on quartz-sericite (phyllic) alteration
(Azadi et al. 2014) 67

Fig. 2. 9. Photomicrographs of propylitic alteration zone in the Kahang porphyry deposit:

a) plagioclase phenocrysts replaced by an aggregate of chlorite (Chl), epidote (Epi) and
calcite (Cal) in the propylitic alteration zone (Afshooni et al., 2013), and b) propylitic
alteration with pervasive epidote and chlorite 68

Fig. 2. 10. The locations of drill cores with lithological units within the Kahang deposit
and its 3D surface topography 69

Fig. 2. 11. Precision result of Cu analysis by Thompson diagram. Plots of replicate

analyses from Cu samples in the Kahang deposit. The mean of the replicate pairs was
plotted along the X-axis, the absolute difference of the two results along the
Y-axis 71

Fig. 2. 12. a) 3D lithology model of the Kahang eastern part (Yasrebi et al., 2012), and
b) lithological fence diagram (See abbreviation list for more details) 78

Fig. 2. 13. a) Alteration model of the Kahang eastern part (Yasrebi et al., 2012), and b)
alteration fence diagram (See abbreviation list for more details) 80

Fig. 2. 14. a) 3D ore-type zonation model, b) fence diagram of ore-type zonation model,
c) 3D dominant ore minerals, d); fence diagram of dominant ore minerals, e)

10
chalcopyrite in stockwork copper mineralisation from hypogene zone, and f) pyrite (Py)
and molybdenite (Mol) mineralisation in hypogene zone (See abbreviation list for more
details) 83

Fig. 3. 1. Self-similarity in a triangle 87

Fig. 3. 2. Changes in dimensions of a fractal shape of Koch Curve

(Zhu et al., 2003) 88

Fig. 3. 3. Histograms for data from the Kahang deposit: a) Cu wt.%, and
b) Mo ppm 91

Fig. 3. 4. 3D maps for original datasets: a) Cu wt.%, b) Mo ppm, c) density t/m3, and d)
RQD % 95

Fig. 3. 5. a) 2D, b) 3D, and c) Google Earth maps of the grid drilling in the Kahang
deposit 100

Fig. 3. 6. Boreholes location (ignoring the three isolated boreholes in the NW) map in
the Kahang deposit with selected closest borehole pairs 101

Fig. 3. 7. Cu histogram based on declustered data 103

Fig. 3. 8. 2D topographical surface of the Kahang deposit 106

Fig. 3. 9. Correlation between borehole data collar heights and topographical

surface 108

Fig. 3. 10. Scatterplots for correlation between Cu (wt.%) and coordinates: a) Cu values
trend in X, b) Cu values trend in Y, and c) Cu values trend in Z 115

Fig. 3. 11. Experimental and theoretical variograms: a) Cu, and b) Mo 117

Fig. 3. 12. Variogram maps for a) Cu in plan 1730 m, b) Mo in plan 1730 m, c) Cu in E-

W section with Northing = 3644585, d) Mo in E-W section with Northing = 3644585, e)
Cu in N-S section with Easting = 638325, and f) Mo in N-S section with Easting =
638325 120

Fig. 3. 13. Project dimensions of the studied area 122

Fig. 3. 14. Steps of IDWAM run in RockWorksTM 15 122

11
Fig. 3. 15. Block models in the Kahang deposit: a) Cu, and b) Mo 124

Fig. 3. 16. Distribution models in the Kahang deposit: a) Cu ≥ 0.4 wt.%, and b) Mo ≥ 200
ppm 125

Fig. 3. 17. Histograms of estimated element concentrations: a) Cu, and b) Mo 125

Fig. 3. 18. a) Cu (%) Grade-Tonnage, and b) Cu (%) average-cut-off diagrams for

Kahang deposit 126

Fig. 3. 19. Cu and Mo plans in the Kahang deposit 131

Fig. 3. 20. Samples (black discs) within Cu values higher than: a) 0.07 wt.%, b) 0.15
wt.% and c) 0.25 wt.% from boreholes carried out in the deposit 132

Fig. 3. 21. Optical correlation (visual verification) between Cu values of borehole and
block model 135

Fig. 3. 22. Correlation chart between original and estimated data using jackknife
resampling 136

Fig. 3. 23. C-V log-log plots: a) Cu, and b) Mo 146

Fig. 3. 24. Geological zones (Cu distribution) including supergene enrichment (a) and
hypogene (c) with modified zonation models via C-V showing regions of supergene
enrichment (b), hypogene (d), main hypogene (e) and enriched hypogene (f) 148

Fig. 3. 25. Histograms of estimated Cu values within the: a) supergene enrichment, and
b) hypogene zones 152

Fig. 3. 26. Mo distribution in supergene enrichment zone (a), hypogene zone based on
Mo C-V model (b), hypogene with Mo > 100 ppm (c), hypogene with Mo > 316 ppm (d)
and Mo enriched zone (e) 153

Fig. 3. 27. Correlation between chalcocite (a), chalcopyrite (b) and chalcopyrite ≥ 0.42
wt.% Cu (e) zones with supergene enrichment zone (c) and main hypogene zone (d)
based on C-V model 154

Fig. 4. 1. Location of boreholes sampled for density and RQD 167

Fig. 4. 2. Density histogram based on raw data for the Kahang porphyry deposit 168

12
Fig. 4. 3. RQD histogram based on raw data for the Kahang porphyry deposit 168

Fig. 4. 4. Experimental and theoretical variogram for RQD 170

Fig. 4. 5. Density block model using estimated data 171

Fig. 4. 6. RQD block model in Kahang porphyry deposit determined using estimated
data 171

Fig. 4. 7. D–V log–log plot in the Kahang deposit 173

Fig. 4. 8. Voxels with density ≥2.7 t/m3 within lithological units: a) porphyritic quartz
diorite, b) andesite, and c) dacite 175

Fig. 4. 9. RQD-V log–log plot in the Kahang deposit 178

Fig. 4. 10. RQD populations within the Kahang deposit based on thresholds defined
from the RQD-V fractal model: a) very poor zones, b) very poor zones, c) poor, fair and
good zones, and d) excellent zones 180

Fig. 4. 11. a) 3D lithology model for porphyritic quartz diorite , b) porphyritic quartz
diorite unit based on the RQD-V model for RQD > 89.12%, c) 3D lithology model for
andesite , d) andesite unit associated with RQD > 89.12%, e) 3D lithology model for
dacite , and f) dacite unit associated with RQD > 89.12% 183

Fig. 4. 12. Correlation between RQD > 70% with density > 2.7t/m3 block model within
the porphyritic quartz diorite 184

Fig. 5. 1. a) Design procedure in an open pit mine with regard to ultimate pit limit
determination (Akbari et al., 2008) and b) steps in mining design and planning by
circular and interdependent analysis (Osanloo et al., 2008b) 192

Fig. 5. 2. a) Estimated Cu block model and b) estimated Cu block model excluding Cu ≤

0.075 wt.%, generated using the C-V fractal model 194

Fig. 5. 3. Land topographical surface of the deposit for optimisation study 195

Fig. 5. 4. Stable pit slopes in the Kahang deposit, input into NPV Scheduler (See also
chapter 4) 197

13
Fig. 5. 5. Density distribution for block models in the Kahang deposit for a) PQD, b)
ANS, and c) DAC lithological units (See abbreviation table for lithological units) 201

Fig. 5. 6. Metal commodity prices at the time of optimisation study: a) copper, and b)
molybdenum (London Metal Exchange, 2015a and b) 203

Fig. 5. 7. Annual discount rates for choice of parameters and model outputs for
domestic and foreign decision makers, a) Iran b) United Kingdom (Trading Economics,
2015) 206

Fig. 5. 8. Kahang pit limit 3D view without consideration of ramps and

safety berms 210

Fig. 5. 9. Pit optimisation within the internal pit shells, stages of 80%, 90% and 100%,
driven by NPV Scheduler based on Table. 5. 8 (Cumulative profit, Incremental revenue,
Incremental total ore and Cumulative total ore) 213

Fig. 5. 10. Comparison between incremental and cumulative NPV values for the Kahang
deposit, driven by NPV Scheduler based on Table. H. 1 (the black arrow indicates
sequence No. 92 which specifies the Kahang ultimate pit limit) 216

Fig. 5. 11. Disposition of boreholes in the Kahang deposit 217

Fig. 5. 12. 3D Cu block model excluding the three isolated boreholes 218

Fig. 5. 13. Cu histogram from original data in the Kahang deposit excluding the isolated
boreholes 218

Fig. 5. 14. C-V log-log plot for Cu concentration with respect to excluding the three
isolated boreholes 219

Fig. 5. 15. Estimated Cu histogram, ignoring 3 boreholes located in the NW part of the
deposit 220

Fig. 5. 16. Cumulative NPV and profit values for the comparative case study driven by
NPV Scheduler (the black arrow indicates sequence No. 90 which specifies the Kahang
ultimate pit limit) 223

Fig. 6. 1. Linear relationship between the NPV and metal prices 229

14
Fig. 6. 2. Present Value (PV) block model for the Kahang Cu-Mo porphyry deposit (the
grey platform distinguishes the boundary between open pit and underground mining
surfaces based on chapter 5, section 5.10.1 and Table. 5. 7) 230

Fig. 6. 3. a) Profit value histogram based on PV block model, and b) NPV histogram
based on the mining sequences for the Kahang deposit 232

Fig. 6. 4. PV log-log plot 233

Fig. 6. 5. PV distribution within the deposit based on the PV-V fractal modelling
consisting of a) moderate population, b) high PVs, and c) extreme population 235

Fig. 6. 6. Voxels with high and extreme PV values within the deposit (the grey platform
distinguishes the boundary between open pit and underground mining surfaces based
on chapter 5, section 5.10.1 and Table. 5. 7) 236

Fig. 6. 7. PV plan views based on the PV-V fractal model in elevations of: a) 2230 m, b)
2240 m, c) 2250 m, d) 2260 m, e) 2270 m, f) 2280 m, g) 2290 m, and h) 2300 m 244

Fig. 6. 8. The NPV-CTO log-log plot in the Kahang deposit 247

15
 List of Tables

Table. 3. 1. Statistical characteristics for Cu and Mo 91

Table. 3. 2. Density analysis from 11 boreholes in the Kahang deposit (See abbreviation
list for rock type) 92

Table. 3. 3. Distribution number of RQD samples among drill cores in the Kahang
deposit 93

Table. 3. 4. The particulars of the selected borehole pairs in the Kahang deposit 104

Table. 3. 5. Variation of voxel size based on mean, median and MAD 104

Table. 3. 6. Voxel numbers and Standard deviations and averages of Cu for different
block models 106

Table. 3. 7. Statistical characteristics for Cu and Mo estimated models 124

Table. 3. 8. Cu and Mo thresholds defined by the C-V model in the

Kahang deposit 145

Table. 3. 9. Matrix for comparing performance of fractal modelling results with

geological model. A, B, C, and D represents the number of voxels overlapping between
classes in the binary geological model and the binary results from fractal models
(Carranza, 2011) 156

Table. 3. 10. Overall accuracy (OA) with respect to hypogene zone as delineated in the
geological model and Cu and Mo main mineralised zones obtained through C-V fractal
model (Values are the number of voxels) 157

Table. 3. 11. Overall accuracy (OA) with respect to potassic and phyllic alteration zones
and Cu main mineralised zones obtained through C-V fractal model (Values are the
number of voxels) 158

Table. 3. 12. Overall accuracy (OA) with respect to potassic and phyllic alteration zones
and Mo main mineralised zones obtained through C-V fractal model (Values are the
number of voxels) 159

16
Table. 4. 1. Classification of Rock Quality Designation, Deere and Miller rock
classification (1966) 166

Table. 4. 2. Overall accuracy (OA), Type I and Type II errors (T1E and T2E,
respectively) with respect to PQD rocks resulted from geological model and high density
rocks obtained through D–V fractal modelling of density data (the obtained values are
the overlap number of voxels between two binary geological and fractal models) 176

Table. 4. 3. Overall accuracy (OA), Type I and Type II errors (T1E and T2E,
respectively) with respect to andesite rocks resulted from geological model and high
density rocks obtained through D–V fractal modelling of density data (the obtained
values are the overlap number of voxels between two binary geological and fractal
models) 176

Table. 4. 4. Overall accuracy (OA), Type I and Type II errors (T1E and T2E,
respectively) with respect to dacite rocks resulted from geological model and high
density rocks obtained through D–V fractal modelling of density data (the obtained
values are the overlap number of voxels between two binary geological and fractal
models) 177

Table. 4. 5. RQD populations (zones) based on three thresholds defined from RQD-V
fractal model 179

Table. 4. 6. Amount of the excellent RQD populations (voxels) defined from the RQD-V
fractal model (RQD ≥ 89.12t/m3) in each major lithological unit within the Kahang
deposit 181

Table. 5. 1. Ore density average of the mineralisation zones within the deposit 199

Table. 5. 2. Prices and mining costs for the Kahang optimisation study 204

Table. 5. 3. COG specification for optimisation study 208

Table. 5. 4. Initial imported data into the optimisation software based on Fig. 3. 13 and
Table. 5. 2 209

Table. 5. 5. Economic specification of the Kahang deposit driven by

NPV Scheduler 209

17
Table. 5. 6. Ultimate pit statistics determined by NPV Scheduler 211

Table. 5. 7. Pit limit reserve in the Kahang deposit 211

Table. 5. 8. Pit optimisation phases of the Kahang deposit 213

Table. 5. 9. Cu thresholds defined by the C-V model in the Kahang deposit, ignoring the
three isolated boreholes 219

Table. 5. 10. Comparison between results obtained from the two Cu block
models 221

Table. 5. 11. Pit limit reserve of the comparative case study 221

Table. 5. 12. Ultimate pit characteristic for the comparative case study 222

Table. 5. 13. Pit optimisation phases for the comparative case study 223

Table. 5. 14. Differences between ultimate pit limits characteristics of the complete
dataset and without the three remote drillholes 224

Table. 5. 15. Differences between optimal extraction sequences characteristics of the

complete dataset (sequence No. 92) and without the three remote drillholes (sequence
No. 90) 225

Table. 6. 1. PV thresholds defined by PV-V model in the Kahang deposit 235

Table. 6. 2. Comparison between OESs regarding cumulative NPV and ore, calculated
from NPV Scheduler and NPV-CTO fractal model 246

18
 List of Appendices

Fig. A. 1. Letter issued by the Kahang deposit project manager authorising the use of
data and choice of input parameters to the pit optimisation study 299

Fig. E. 1. Density plan view for layer # 32 (Z =1900 m) 310

Fig. E. 2. RQD plan view for layer # 32 (Z =1900 m) 311

Fig. E. 3. Density plan view for layer # 37 (Z =1950 m) 312

Fig. E. 4. RQD plan view for layer # 37 (Z =1950 m) 313

Fig. E. 5. Density plan view for layer # 42 (Z =2000 m) 314

Fig. E. 6. RQD plan view for layer # 42 (Z = 2000m) 315

Fig. F. 1. 2D economic model (Section level 1970 m, Plan view) 318

Fig. F. 2. 2D economic model (Section level 1900 m, Plan View) 319

Fig. F. 3. 2D economic model (Section 638400 E North-South View) 320

Fig. F. 4. 2D economic model (Section 3644688.00 N East-West View) 321

Fig. G. 1. 2D pit limit (Section level 2225 m, Plan View) 322

Fig. G. 2. 3D Ultimate pit limit view of the Kahang deposit including all boreholes
(Section 5140.00 E) 323

Table. B. 1. 399 randomised samples for Cu selected and analysed for assay quality
assurance and quality control 300

Table. C. 1. Fisher distribution F(n1,n2) with n1 and n2 degrees of freedom, α = 0.025

and 97. 5% of confidence level (Emery, 2012) 308

Table. D. 1. Critical values for student’s T distributions (column headings denote

probabilities’ α above tabulated values: Emery, 2012) 309

Table. F. 1. The individual economic properties at all elevations (generated by NPV

Scheduler software) 316

Table. H. 1. Pit optimisation: Extraction sequences for incremental and cumulative NPV
generated by NPV Scheduler software (sequence No. 92 specifies OES) 324

19
List of Abbreviations

NPV: Net Present Value

C-V: Concentration-Volume
D-V: Density-Volume
RQD: Rock Quality Designation
RQD-V: RQD-Volume
PV-V: Present Value-Volume
NICICO: National Iranian Copper Industries Co
XRD: X-ray Diffraction
EPMA: Electron Probe Micro Analyser
SEM: Scanning Electron Microscopy
PIMA: Portable Infrared Mineral Analyser
PV: Present Values
IDW: Inverse distance weighted
S-A: Spectrum-Area
C-A: Concentration–Area
C-D: Concentration-Distance
PS: Positive strong
PW: Positive weak
MW: Minus weak
OES: Optimal extraction sequences
IP: Induced polarization
RS: Resistivity sounding
Py: Pyrite
Ccp: Chalcopyrite
Mol: Molybdenite
Py I: Early pyrite
Py II: Late pyrite
VPR: Volcano-plutonic rocks
Bt: Biotite
S-Bt: Secondary biotite

20
Qtz: Quartz
Plg: Plagioclase
Op: Opaque minerals
Kf: K-feldspar
Jar: Jarosite
Epi: Epidote
ICP-MS: Inductively coupled plasma mass spectrometry
OK: Ordinary kriging
MAD: Median Absolute Deviation
SD: Standard deviation
IDWAM: Inverse Distance Weighted Anisotropic method
SK: Simple Kriging
G-T: Grade-Tonnage
T1E: Type I error
T2E: Type II error
OA: Overall accuracy
IBSD: In-situ block size distribution
V. RQD: Volumetric Rock Quality Designation
PQD: Porphyric quartz diorite
COG: Cut-off grade
UPL: Ultimate Pit Limit
BESR: Break-Even Stripping Ration
ODBC: Open Database Connectivity
CAF: Cost Adjustment Factor
CBI: Central Bank of Iran
WTO: World Trade Organisation
DCF: Discounted cash flow
NPV-CTO: NPV-Cumulative Total Ore

21
List of abbreviations for lithological units, hydrothermal alterations, ore minerals and ore-type
zonation in the Kahang deposit

22
 Glossary of Notations

F: F-distribution or continuous probability distribution

S1 and S2: Variances for pair of samples

n1 and n2: Number of observations

α: Probability value

H0: Null hypothesis

H1: Alternative hypothesis

σ1 and σ2: Variances of populations

n: Degrees of freedom

𝑑̅ : Mean difference of the pair samples in terms of their grades

Sd: Differences of standard deviation

SE: Standard error of the mean difference

𝑛̅: Number of paired samples

T: T-test statistic

e: Error analysis

Xi and Yi: Measured values for duplicated samples

U: Interpolated value

x: An interpolated unknown point

xi: An interpolating known point

D: A given distance

N: Total number of known points used in interpolation

P: Positive real number called the power parameter

ɣ (h): Theoretical variogram

h: Lag separate distance

23
q: Partition function

: 2D or 3D blocks dimensions

: A parameter in investigation

q: value of  at any given moment

A: Area

ρ: Concentration, density, RQD, PV values

υ: Threshold

a1 and a2: Characteristic exponents

Ah and Av: Areas horizontal (Plans) and vertical (cross sections) directions

Ah(ρ≤υ), Ah(ρ≥υ), Av(ρ≤υ) and Av(ρ≥υ): Areas with concentration values smaller and values
greater than the contour value ρ

h and v: Horizontal and vertical section

𝑓(𝛼 ): Continuous function

(q): Auxiliary function

á = áMax or á = áMin: Fixed values

𝑉(𝜀): Containing volume

C: constant value

𝑉(𝜌): Occupied volume

V(ρ≤υ) and V(ρ≥υ): Occupied volume with values that is smaller or greater than contour
value ρ

D: Fractal dimension

Jv: Volumetric Discontinuity Count for RQD

I1: Income of Cu per each voxel

I2: Income of Cu and Mo per each voxel

24
MC: Mining cost for a voxel

MR: Mining recovery

PC1: Processing cost for Cu

PC2: Processing cost for Mo

PCu: Metal commodity price for Cu

PMo: Metal commodity price for Mo

PR: Fraction recovery

D: Voxel density

V: Voxel volume

MCPU: Mining cost per volume unit

PCPU: Processing cost per tonnage unit

G1: Cu concentration values

G2: Mo concentration values

V(ρPV≤υ) and V(ρPV≥υ): Volumes (V) with PV values that are smaller and greater than
PV threshold values

ρPV: PV values

CTO(ρNPV≤υ) and CTO(ρNPV≥υ): Cumulative total ore with NPV values that are smaller
and greater values than NPV threshold values

ρNPV: NPV values

25
CHAPTER ONE. Introduction

26
1.1 Problem Description

Mineral excavation by open pit mining methods requires huge investment which will

inevitably rise over the life of a mine due to increases in the amount of cumulative waste

materials and mining costs (e.g., See Appendix. H for cumulative data; Caccetta and

Giannini, 1988; Hustrulid and Kuchta, 2006; Akbari et al., 2008; Elkington and Durham

2011). Before starting the mining operation, it is necessary to design the final shape and

size of the pit in order to determine the minable reserve and amount of waste to be

removed. Following this, an optimised block model should be produced showing ore

grades, density and Rock Quality Designation (RQD). RQD is the method perhaps most

commonly used for characterising the degree of jointing in drill cores and can be

considered as an expression of intact core lengths greater than a threshold value of 0.1

m along any bore hole. An increase in the number of joints in a rock mass causes a

decrease in RQD (Bieniawski, 1984) and Net Present Value (NPV), the latter which is

defined as the sum of all cash flows discounted to a specific time in an investor’s minimum

rate of return, or discount rate. NPV is a measure of value created by investing in a project

(a mining project in this scenario) and not investing capital in any other project at the

minimum rate of return. NPV higher than zero is acceptable however; an NPV equal to

zero is a breakeven. The cumulative NPV indicates the value of one additional year of

cash flow and its impact on the overall project NPV (Stermole and Stermole, 2012).

Calculation of the final pit limit, which this study aims to optimise, is a function of numerous

variables, especially NPV, and may be re-evaluated many times over the mine life (Lerch

and Grossmann, 1965; Dowd and Onur, 1992; Akaike and Dagdelen, 1999; Hustrulid and

Kuchta, 2006).

27
Separation of different populations based on ore grades and consequently the ID of

mineralised zones in geological modelling for excavation of minable ores, specifically

identifying ‘barren’ host rock from the main ore body, is one of the fundamental issues

within a mining operation. Conventional methods for characterising mineral assemblages

(e.g. X-ray Diffraction (XRD), Electron Probe Micro Analyser (EPMA), Scanning Electron

Microscopy (SEM) and Portable Infrared Mineral Analyser (PIMA)) have been used since

the 1960s to delineate mineralised zones however these do not have enough detailed

information based on ore grades particularly in the porphyry deposits due to variation of

ore grade distribution within block models (Schwartz, 1947; Lowell and Guilbert, 1970;

Cox and Singer, 1986; Sillitoe, 1997; Beane, 1982; Berger et al., 2008).

Fractal geometry has a distinctive power to distinguish natural populations (zones) within

orebodies. The research within this thesis utilises the Concentration-Volume (C-V) fractal

model developed by Afzal et al. (2011) to delineate mineralised zones in terms of grades

to obtain an optimised orebody model. Data from rock mass characterisation is then input

into newly developed Density-Volume and RQD-Volume fractal models to determine an

ultimate pit limit for a maximum NPV.

The importance of zone separation in porphyry ore deposits is as follows:

1. The key property of porphyry ore deposits is their low ore grades and high tonnage.

However, the importance of this issue should be attributed to this fact that the different

mineralised zones (leached, oxidised, supergene and hypogene) of those deposits are to

28
be distinguished carefully in order to demonstrate a comprehensive mine design and

planning (Sim et al., 1999; Cheng and Agterberg, 2009; Sadeghi et al., 2012). For

example, supergene enrichment zones are generally accepted to be the primary target

when mining porphyry deposits (Hartley and Rice, 2005; Berger et al., 2008; Asadi et al.,

2015). They are enriched in ore elements especially copper. The supergene enrichment

zone consequently has high values in terms of money pay-back due to high grades. It is

also usually located near surface which can reduce the costs of mining (Alpers and

Brimhall, 1989; Sillitoe, 2005). In addition, the concentrator capacity is usually determined

during the pre-feasibility study so it is necessary that the various zones with

corresponding different grades and tonnage be identified to regulate the concentrator

(Hustrulid and Kuchta, 2006). Detailed theoretical and experimental investigations of

porphyry ore deposits have provided us with a good knowledge of how they were formed.

It is now relatively easy to establish natural geochemical variability in terms of ore grades.

However, geoscientists are now able to decipher the reasons for extreme variability in

element concentrations using mathematical-oriented practices such as fractal modelling.

This type of approach would continue to enhance the potential to identify relations

between ore grade and their spatial distribution within a deposit (Monecke et al., 2005;

Davies and Mundalamo, 2010; Spalla et al., 2010; Sadeghi et al., 2012).

2. The hypogene zone, which often comprises the bulk of a porphyry Cu deposit, is

generally located below the supergene enrichment zone and contains lower grades with

high tonnages of ore. Therefore, the location of this zone must be accurately determined

29
because huge amount of feed input to the processing plant is generally from this zone

(Atapour and Aftabi, 2007).

3. The oxidised zone, which overlies the enriched supergene blanket, contains generally

low but variable levels of Cu. By the end of the 20th century, this zone was generally

treated as waste but after developing leaching and bioleaching methodologies, Cu has

been exploited from this zone (e.g., copper (I) oxide or cuprous oxide for example cuprite

(Cu2O), copper (II) oxide or cupric oxide for instance tenorite (CuO), copper carbonates

such as azurite and malachite, copper sulphate mineral such as chalcanthite, copper

silicates for example chrysocolla and dioptase). Therefore, determination of the

boundaries for this zone is fundamentally important so as to specify the distinct section

exposed to leaching.

4. The discrimination of ‘barren’ host rock from the orebody and delineation of the zone’s

boundaries play a significant role in determining an optimised block model to be examined

for final pit limit and correspondingly pit optimisation. Unfortunately, in most cases the

‘barren’ host rock is mistakenly identified as ore and consequently the size of the reserve

is exaggerated (Agterberg et al., 1993; Panahi et al., 2004; Zuo et al., 2012; Zuo and

Wang, 2015). This results in egregious errors during production planning. On the other

hand if ‘barren’ rocks and ore are not properly delineated, some parts of the ore deposit

may be lost. As a result, an exact determination of the boundary and the locations of the

zones is necessary for the long-term planning of Cu mines and consequently should not

be subject to major changes during mining. For example, if basic requirements are not

30
met, the formerly selected input feed grade (considering plant initial ore grade) to the

processing plant will be lowered significantly which will reduce production. This is typically

the case in the large copper mines of Chile and the USA (Carrasco et al., 2004; Parhizkar

et al., 2011 and 2012).

The use of geological data including structure, lithology and mineralogy, main ore types

and associated secondary elements, geophysical and geochemical anomalies as well as

morphology of the ore deposit is considered to be the most important method for

separating the various mineralised zones in descriptive models which were proposed and

developed by Cox and Singer (1986). These models have major disadvantages as

follows:

1. Geological core-logging is subjective rather than quantitative. In the event that both

thin and polished sections of all cores are prepared to improve the accuracy, this process

is both costly and time-consuming.

2. The grade of the ore element (particularly Cu) is not observed (visual assessment) with

these methods while the variation of the grades in each zone is an obvious and salient

feature which has to be always measured in a laboratory. Fractal modelling has proved

their superiority to the classical statistical and conventional geological methods as follows:

 In classical statistics, for the purpose of determining the boundaries in mineralised

zones, frequency distribution of a related ore element in an intended area must

31
 adhere to a normal distribution. In addition, separation of different populations

based on mean and standard deviation should be carried out with normalised data.

This requirement is not always met in data. In addition, local neighbourhood

statistics can provide less statistical information which is less biased than that of

global statistics, such as mean and SD, because geochemical data generally

satisfy non-normal distributions and contain outliers. However, there is no need to

normalise data when the distribution of elemental concentrations is determined

through fractal modelling (Agterberg et al., 1993; Cheng et al., 1994; Agterberg et

al., 1996; Sim et al., 1999; Zhang et al., 2007).

 Traditional methods consider only the frequency distribution of the elemental

concentration and ignore its spatial variability. Specifically, the information about

the spatial correlation is not always available. Moreover, statistical methods e.g.,

by histogram analysis or Q–Q plots assume normality or log-normality and do not

respect the shape, extent and magnitude of a studied area (Armstrong and

Boufassa, 1988; Clark, 1999; Reimann and Filzmoser, 2000; Limpert et al., 2001).

A power law relationship between quality parameters (e.g., ore element) and their

occupied spatial positions (e.g., volume, area and perimeter) are illustrated in the

fractal/multifractal modelling to solve the problematic issues mentioned above

(Rafiee and Vinches, 2008; Sadeghi et al., 2015).

 Those values which are not within the range (outliers) must be identified and

eliminated accordingly; otherwise they lead to the intended study having unreal

32
 results. All data are contributed to fractal/multifractal modelling which is help for

better separation of different mineralised zones (Xiao et al., 2014; Zuo, 2014).

Given the problems as mentioned above, using a series of newly established methods

based on mathematical analyses seems to be inevitable. By using these methods, one is

able to delineate the various mineralised zones in terms of grade and therefore the

accuracy of one’s pit limit optimisation. As a result, a C-V fractal modelling technique is

proposed as a better method in order to identify geochemical zones, rock mass

characteristics and economical populations in the Kahang Cu-Mo Porphyry deposit. The

deposit is identified as an important Cu-Mo and deposit located approximately 73 km NE

of Isfahan in Central Iran. It contains more than 100 million tonnes of sulphide ore with an

average grade of 0.5 wt.% Cu and 90 ppm Mo. This deposit occurred within the Cenozoic

Urumieh–Dokhtar magmatic belt, one of the subdivisions of the Zagros Orogenic Belt.

This belt extends for some 2,000 km from NW to SE Iran. Many of the Iranian large

porphyry Cu deposit such as Sarcheshmeh, Sungun and Meiduk are situated within this

belt. Geological, geophysical, geochemical, alteration patterns as well as drilling data

show that there could be a large Cu porphyry deposit at Kahang.

1.2 Objectives of the Research

The most important issue of this PhD thesis is to attempt to develop an innovative method

to separate oxidised, supergene-enriched and hypogene zones from ‘barren’ host rock in

a Cu-Mo porphyry deposit on the basis of fractal geometry using geochemical data. As a

33
result, an approach called C-V fractal modelling has been developed to determine proxies

from grades, densities, RQDs and Present Values (PVs) within a deposit for pit limits.

The general objectives of this study are as follows:

 To create a 3D geological model for lithology, alteration, zonation and

mineralisation

 Generate a grade block model via the C-V model

 Rock mass characterisation using fractal/multifractal modelling

 Determine economic principals

 Calculate final pit limit for the Kahang deposit

 Create an economic block model from Present Value-Volume (PV-V) fractal

modelling

 Calculate pushbacks resulting from the PV-V fractal model

1.3 Application of Fractal/Multifractal Modelling with Reliance on

Geochemical Population

Euclidian geometry identifies geometrical shapes with an integer dimension say 1D, 2D,

3D, etc. However, there are many other shapes amongst spatial objects, whose

dimensions cannot be mathematically described by integers but by real numbers or

fractions (Bölviken et al., 1992; Agterberg et al., 1996; Aghanabati, 2004; Ali et al., 2007).

These spatial objects are called fractals. In an abstract form, fractals describe complexity

in data distribution by their fractal dimensions. A wide range of complex

structures/features and geological phenomena of interest to geologists and geochemists

have been quantitatively recognised using fractal/multifractal modelling over the past

34
several decades, mainly how to identify geochemical populations and quantify the spatial

distribution of geochemical data. Various geochemical processes can be described based

on differences in fractal dimensions obtained from analysis of relevant geochemical data.

Recognition of geochemical populations is a crucial aspect for applied geochemists to

effectively detect geochemical populations from background (Darnley et al., 1995; Plant

et al., 2001; Lima et al., 2003, 2005 and 2008; Albanese et al., 2007).

Ore elements, especially trace elements, do not follow a normal or lognormal distribution.

However, they follow a positively skewed distribution toward high values (Ahrens, 1954

and 1957; Krige, 1966; Turcotte, 1986; Reimann and Filzmoser, 2000; Agterberg, 2007;

Carranza, 2009). Recent investigations of geochemical features have shown that self-

similarity or self-affinity are significant properties of geochemical data (Bölviken et al.,

1992; Cheng et al., 1994; Zuo et al., 2009a and b; Afzal et al., 2011; Zuo and Wang,

2015). The most effective way to distinguish geochemical anomalies from the background

is to suggest a comprehensive technique which can be mathematically interpreted. The

typical and most widely used method for detection of geochemical anomalies is the setting

of threshold values which include upper and lower limits of background variations

(Hawkes and Webb, 1962; Cheng et al., 1994; Xu and Cheng, 2001; Li et al., 2003; Lima

et al., 2003; Afzal et al., 2010, 2011and 2012; Agterberg, 2012; Zuo et al., 2015).

However, conventional geological methods, exploratory data analysis and multivariate

statistics are based on the frequency distribution of geochemical values and neglect

spatial variation (Tukey, 1977; Behrens, 1997; Yousefi et al., 2012 and 2014). In addition,

exploration geochemical data are typically spatially dependent and therefore a couple of

35
frequency–space-based methods such as the inverse distance weighted (IDW) and

different kriging methods have been used (Lam, 1983; Zimmerman et al., 1999). Although

these methods consider the spatial distribution of elemental concentrations, they do not

consider that spatial variability is rugged and singular rather than smooth and

differentiable. The main attraction of fractal/multifractal theory is its capability to quantify

irregular and complex phenomena or processes that exhibit similarity over a wide range

of scales; this is regarded as self-similarity (Mandelbrot, 1983; Zuo and Wang, 2015).

Fractal theory, which was developed by Mandelbrot (1983), has been widely applied in

the geosciences sector since the 1980s and up to the present (e.g., Turcotte, 1986;

Agterberg et al., 1993; Cheng et al., 1994; Sim et al., 1999; Goncalves et al., 2001; Shen

and Zhao, 2002; Ali et al., 2007; Yasrebi et al., 2013a). Methods of fractal analysis also

serve to illustrate relationships between geological, geochemical and mineralogical data

and spatial information derived from analysis of mineral deposit occurrence data

(Carranza, 2008; Carranza et al., 2009; Goncalves et al., 2001). A good understanding

of geological and geochemical controls on mineralisation is essential in the recognition

and classification of geochemical populations based on methods of fractal analysis which

indicate relations between ore grade and their spatial distribution within a block model

(Cheng, 1999; Sim et al., 1999; Li et al., 2003; Carranza and Sadeghi, 2010). Fractals

are characterised by a scaling law that relates two variables: the scale factor and the

physical properties of the object being measured. This scaling relationship is described

by a power law function, which in turn describes the inherent physical attributes of the

object being analysed (Takayasu, 1990; Lauwerier, 1991; Ortega et al., 2006). The

36
exponent of the power law function refers to the fractal dimension. Fractal dimensions in

geological and geochemical processes correspond to variations in physical attributes

such as rock type, nature of the hydrothermal fluids and alteration, structural features and

dominant mineralogy, and so on (Sim et al., 1999; Cheng, 2007; Cheng and Agterberg,

2009; Afzal et al., 2013a and 2014; Yasrebi et al., 2013a, b and 2014). Therefore, fractal

dimensions of variations in geochemical data can provide useful information and

applicable criteria to recognise and classify mineralised and ‘barren’ zones within a study

area. Various log–log plots in fractal methods are considered to be useful tools in

separating geological populations. Classification of geochemical data within threshold

values can be recognised and determined to indicate breakpoints within these plots.

These geochemical threshold values recognised via fractal analysis are usually correlated

with geological field information. Multifractal theory is used as a theoretical framework to

explain the power–law relationship between areas enclosing concentrations below a

given value and the actual concentrations themselves (Halsey et al., 1986; Evertz and

Mandelbrot, 1992).

The fractal method has several limitations and accuracy issues, especially when the

boundary effects on irregular geometrical data sets are involved (Agterberg et al., 1996;

Goncalves, 2001). The Concentration-Area (C-A) method (Evertz and

Mandelbrot, 1992; Cheng et al., 1994), which is the basis of the C-V fractal model, seems

to be equally applicable in all cases, which is probably rooted in the fact that geochemical

distributions mostly satisfy the properties of a multifractal function. There is some

evidence that geochemical distributions are fractal in nature and behaviour, at least

37
empirically according to Bölviken et al. (1992). Some approaches seem to support the

idea that geochemical data distributions are multifractal, although this point is far from

proven (Cheng and Agterberg, 1996; Turcotte, 1997; Goncalves, 2001; Afzal et al., 2010,

2013a and b). This idea may help in the development and validation of a method for

elemental geochemical distribution analysis.

1.3.1 Introduction to Common Fractal Models

Cheng et al. (1994) proposed the Concentration-Area (C-A) fractal model, which is used

frequently for modelling geochemical anomalies and discriminating geochemical

anomalies from background, relates the element concentration to the area enclosed by

concentration contours by a power law relation (Carranza, 2009; Zuo et al., 2012). He

applied the C–A fractal model to lithogeochemical data of the Mitchell-Sulphurets

precious metal district, British Columbia and found that various fractal patterns exist inside

and outside the potassic, sulfidic, and silicic alteration zones. The Spectrum-Area fractal

model (S–A), which was proposed by Cheng et al. (1999), is a version of the C–A

(Concentration–Area) model which separates overlapping populations using more than

one cut-off value. Li et al. (2003) introduced the Concentration–Distance (C–D) fractal

model for discriminating geochemical anomalies from background. These models have

been widely used to identify anomalies as well as for determining the geochemical

baseline in environmental studies (e.g., Cheng et al., 1994; Cheng and Agterberg, 1996;

Cheng, 1999; Gonçalves et al., 2001; Xu and Cheng, 2001; Li et al., 2003; Panahi et al.,

2004; Cheng, 2007; Albanese et al., 2007; Afzal et al., 2010; Wang et al., 2013; Asadi et

al., 2014 Luz et al., 2014; Sun et al., 2014). Cheng (2012) suggested a Density–Area

38
Power-Law model to systematically confirm that singularity analysis is effective for the

identification of weak geochemical anomalies. Furthermore, Wang et al. (2012)

investigated the fractal relationship between orebody tonnage and thickness by tonnage–

thickness model and metal tonnage–grade thickness model for better understanding

orebody spatial distribution (Zuo and Wang, 2015).

1.4 Methodology
The determination of a final pit limit in the Kahang Cu-Mo porphyry deposit is the major

aim of this research. To achieve this, subsurface data obtained by boreholes is entered

into the RockWorksTM software enabling one to generate a 3D geological model based

on lithology, alteration, mineralisation and zonation. Geostatistical studies were then

conducted in order to build the Cu and Mo block models based on the dataset with

utilisation of SGeMS software. The next step was to test different approaches, namely

Concentration-Volume (C-V), Density-Volume (D-V), RQD-Volume (RQD-V) and Present

Value-Volume (PV-V) fractal models, to delineate various populations in terms of Cu and

Mo values, densities, RQDs and valuable voxels respectively. Consequently, data for the

main ore body of the Kahang, including ore grade, density, tonnage and rock type (ore

and waste), was used to generate a prototype for the determination of internal pit shells,

extraction sequences (Nested pits) and ultimate pit limit, which is a collection of optimised

pits. All pits were calculated on the basis of the Lerch & Grossman algorithm (Hustrulid

and Kuchta, 2006). Indeed, constructing optimised nested pits is considered an ‘art’

applied by design engineers to specify the ultimate optimised pit as well as to determine

the extraction sequence of the blocks, with respect to the obtained thresholds via fractal

39
modelling log-log plots. The result of this study will be a reserve estimation and the

modification of existing mineralised zones with respect to the different ore grades within

the deposit. Finally, NPV Scheduler was employed in order to establish the final pit limit

in terms of the maximum NPV and associated ‘’Nested pits’’ to produce an optimised pit

limit.

1.4.1 Introduction to NPV Scheduler

Since the 1960s, computerised open pit optimisation methods have been used and most

major mining companies employ some form of these methods in the design of their open

pit operations. These computerised optimisations can also be utilised to aid in the

transition from open pit to underground mining methods. Examples of widely used

systems include the NPV Scheduler. The primary focus of these software systems is to

determine an optimum size and shape of open pits to enable the generation of production

schedules. This software encompasses a number of processes which utilise

computerised grade block models and generate detailed economic analysis of different

open pit mines (See chapter 5 of this thesis and Appendices F and H). This analysis,

which includes discounted cash flows, also demonstrates productive information to assist

in the mine planning and scheduling of open pits (NPV – Scheduler, 2001).

1.4.2 Introduction to Lerch and Grossmann Algorithm with Reliance on

Resource Modelling
In the late 1960s, researchers were only focused on the determination of the ultimate pit

limit (Zhao and Kim, 1992; Shishvan and Sattarvand 2015). The Lerch and Grossmann

algorithm, which is based on three-dimensional graph theory, is the most commonly used

40
optimisation algorithm which takes into account the influence of a grade block model,

operating costs, product prices, slope geometry, etc (Lerch and Grossmann, 1965;

Dimitrakopoulos et al., 2007). It is also used in mining optimization software as the

industry standard, for example in NPV Scheduler and Gemcom’s Whittle software (Whittle

1998a, b and 1999), to find the optimal pit and pushbacks. The algorithm uses different

revenue factors to generate a value-based mining sequence strategy to design pit shells

(Dincer 2001; Bastante et al., 2008; Grenon and Hadjigeorgiou, 2010; Shishvan and

Sattarvand 2015). Early pit shells are constructed using high-grade blocks and a low

stripping ratio. The results also consider practical considerations such as haul road

access, cut-off grades and processing, etc. To maximise the use of block modelling

functions and optimise the pit design process, block modelling and slope stability analysis

have to be fully integrated. This is a logical extension to assign mines rock types and

grades to every block .This process will be further optimised by defining every block

location especially those blocks with high value of NPV (e.g., the use of fractal geometry

in this thesis: See chapter 6). The algorithm works as follows:

First, a directed graph (Bondy and Murty, 1976) is produced with the nodes of the

orebody, the blocks in the orebody model. These connected blocks have certain

restrictions, for example precedence and slope limitations. The method produces a tree

regarding Lerch & Grossman algorithm as a set of combined voxels with a dummy node

and strong and weak arcs between the nodes (Fig. 1. 1). When the restrictions are

satisfied, the pit has the maximum closure graph at a scaled capacity (Lerch and

Grossman, 1965; Zhao and Kim 1992; Seymour, 1995; Hustrulid and Kuchta 2006;

41
Meagher et al., 2010). In step one, the blocks/nodes are connected to the dummy node,

X0, with arcs from X0. Step two indicates the initial normalised tree, the positive strong

(PS) arcs are plus arcs supporting blocks with NPVs higher than zero (strong vertices)

and positive weak (PW) arcs indicates blocks with NPVs less than zero (weak vertices)

which have negative significance in calculation of the total open pit mining project NPV.

Step three indicates merging vertices X4 and X6; the arc between X0 and X6 will be

removed out. Minus weak (MW) denotes a minus arc supporting a strong arc. Step four

illustrates the tree when all the weak vertices above X6 are merged. Step five shows the

final graph closure with the strong vertices associated to the dummy node. In total, The

Lerch and Grossmann algorithm is based on two theorems (Caccetta and Giannini, 1986):

1. The maximum closure of a normalised tree is the set of that tree's strong

vertices.

2. A normalised tree can be found such that the set of strong vertices in this tree

constitutes a closure of the graph so the set of strong vertices is the maximum

closure of the graph with the highest NPV.

42
 Fig. 1. 1. An example of the graph closure in the Lerch and Grossman algorithm

(Meagher et al., 2010)

The optimised pit generated by the Lerch and Grossmann algorithm always has a crest

within a studied block model so no produced pit will break through the side of the model.

Consequently, if the region or model area is too small, an underestimated optimised pit

will be resulted (Kim, 1978; Frangois-Bongarcon and Guidal, 1982; Koenigsber, 1982;

Seymour, 1995; Hochbaum and Chen, 2000; Bernabe, 2001; Ramazan, 2007). On the

other hand, if the block model is too large and the optimisation software (NPV Scheduler

used in this thesis) cannot fit into RAM, so the optimisation software will work slowly. To

offset this, engineers wish to bulk the block model voxels together as depicted in Fig. 1.

2.

43
 Fig. 1. 2. Bulking the block model voxels together (Mart and Markey, 2013)

1.4.2.1 Introduction to Other Methods for Mine Planning

A 3D program called GEOVIA Whittle™, introduced by Whittle (1985), was a computer-

based implementation of the Lerch and Grossmann method which used a block model,

whose blocks have economic values representing the net cash flow that result from

mining the block in isolation (Whittle, 1988, 1989 and1999). However, the resulting

optimal pit did not use discounted cash flows.

The Floating Cone method, which is the simplest and fastest technique to determine

optimum ultimate pit limits to which variable slope angle can be easily applied, repeatedly

searches for and checks the total value of block groups forming inverted cones. Total

cones are identified for mining if their total value was positive. This procedure is iterated

until no more positive cones are recognised. However, this method cannot guarantee the

final pit is optimum. Other block groups (as mentioned above) also implemented a two-

and-a-half dimensional Lerch and Grossmann algorithm (Dimitrakopoulos et al., 2002;

Osanloo et al., 2008a; Asad and Dimitrakopoulos, 2013).

44
The 4D (and subsequently Four-X) programs also use the same Lerch and Grossmann

technique to generate a set of nested optimal pits. Each pit that is optimal is used to guide

different mining schedules. Financial analysis of these programs which consider

discounted cash flows allows selection and sensitivity analysis of the best pit (Dowd,

1994; NPV – Scheduler, 2001; Osanloo et al., 2008a; Askari-Nasab et al., 2011).

1.5 Specific Economic and Political Context of Mining in Iran

The mining sector is key to sustainable development in many countries such as Iran

(Sameni Keivani and Khalili Sourkouhi, 2014). The following text describes Iranian

government policy, programs and aims with regard to the mining sector (revealed by

Mr.Nematzadeh, the minister of Industries, Mines and Commerce of Iran, at the Iran

Parliament, 2015).

“Governmental and private mining sectors in Iran are one of the largest and most effective

sectors of the country’s economy and own a vast diversity and complexity compared with

other sectors, providing considerable and noteworthy effect on the economy improvement

of Iran. Mining and in one single word mineral productions, is the motive engine of the

country’s economy which has a crucial role in the economic growth, decrease of inflation,

unemployment and improvement in competition and rivalry.

The general policy of the Iranian government is to set up an economic development to

move towards a position in which it will be able to have a noteworthy position in the world

economy based on mining industries to generate national wealth.

45
Iran by owning 57 billion tonnes of mineral ores (proved and probable), with 69 different

ore minerals, ranks first in the Middle East and is on a par with the top 10 mineral

producers globally. From among 7036 licensed mines, 5060 mines with reserves

amounting for 40 billion tonnes are active and in recent years, around 341 million tonnes

of minerals valued at 3.7 billion US dollars (with an average of each tonne equal to 21.4

dollars) have been extracted and nearly 3 billion US dollars of these have been exported.

Only two percent of the total mines of the country belongs to the governmental sector,

including 25% of total extracted minerals in the country, which is over 35% of minerals

production in terms of economic value. General aims of the Iranian government regarding

mines and mining industries development are as follows:

1- Increase competition within the country’s mining sectors.

2- Increase the value added share of mineral products within the country.

3- Increase the country’s minerals exports.

4- Increase the amount of mineral products with high technology/value added

and consequential exports.

5- Increase the role of the private sector mining activities.

6- Promotion of environmental standards towards access to universal

sustainable development goals.

7- Effort towards joining the World Trade Organisation and utilisation of the

capacities thereof.

46
The Iranian Government’s quantitative goals regarding mining development are:

1- Reach an annual average growth of value added of 12 percent.

2- Export high quality mineral products up to 30 percent of the total country’s

export by the end of tenure of the presidency of Dr. Rohani.

3- Industrial and mineral exports portion compared to total world industrial

exports at the closure of presidency of Dr. Rohani to be up to 3 percent.

4- The ratio of industrial and mineral exports to industrial and mineral imports

at the end of presidency of Dr. Rohani to be equal to 100 percent.

5- Absorption of direct foreign investment for the country’s mining projects

(annual average of 8 billion dollars).

The Iranian Government policies regarding mining development are:

1- Review of strategic documents covering development of industry, mining

and trade.

2- Upgrading the potential of small and medium mineral industries towards

expansion of their products as exports.

3- To facilitate absorption and development of foreign investment.

4- Protection of private mining sectors for renovation and amelioration.

5- To help promote competitiveness.

6- To develop industrial and mining facilities and help restart ceased or

inactive mines.

47
 7- To help develop an optimal consumption management of energy in the

mining sector.

8- Efficient support for research and development in the mining sector.

Here it is worth mentioning that the data sources to conduct this PhD research were

provided by the Kahang deposit owner (Appendix. A), the National Iranian Copper

Industries Co (NICICO) which has numerous responsibilities including extraction and

utilisation of copper mines, production of copper concentrates and manufacturing copper

products such as cathodes, slabs, billets and 8 mm wire rods.

1.6 Organisation of the Thesis

Chapter one gives the background to, and the problem statement for the research

undertaken on the importance of delineation of mineralised zones in a Cu-Mo deposit.

The aims and objectives of the research are presented along with a brief description of

the methodology to achieve the outlined objectives.

Chapter two deals with the geology and associated mineralisation in the Kahang deposit

providing 3D models of lithology, alteration, zonation and mineralisation.

Chapter three introduces the C-V fractal model. Additionally, a correlation between results

achieved from the C-V fractal model and those from geological models is used to optimise

the delineation of mineralised zones. A logratio matrix has been employed to validate the

C-V fractal model for the Cu and Mo main mineralised zones.

48
Chapter four proposes the D–V and RQD-V fractal models to delineate rock

characteristics including density and RQD within the Kahang porphyry deposit. A

correlation of results from the D–V fractal and lithology models was carried out to illustrate

that the main lithological unit is associated with high values of density and also has a

strong correlation with high values of RQD. The log-ratio matrix was employed to validate

the D–V fractal model for density with the main rock type of the deposit. The results reveal

that there is a multifractal pattern of rock characteristics with respect to RQD for the

Kahang deposit.

Chapter five discusses the determination of an ultimate pit limit using the results achieved

from the proposed fractal models in the former chapters. Following this, the NPV

Scheduler was employed in order to establish the final pit limit in terms of the maximum

NPV and associated mining sequences. Finally, a comparative case study was also

conducted by ignoring three isolated boreholes located in the NW part of the deposit.

Chapter six introduces a fractal model to achieve a best mining scenario and strategy for

an earlier pay-back. In addition, a new method is proposed to identify an optimal

extraction sequences (OES).

Chapter seven summarises the main conclusions drawn from the entire research project.

The knowledge gained from each specific investigation is summarised along with the

contributions to knowledge. The chapter concludes with the author’s recommendations

for future work on the topic.

49
CHAPTER TWO. Geology and Associated
Mineralisation

50
2.1 Regional Geology

The Kahang Cu Porphyry deposit is located approximately 73 km from Isfahan, in Central

Iran. This deposit is situated in the central part of the Cenozoic Urumieh-Dokhtar

magmatic belt, which extends for 2000 km and is 150 km wide, from NW to SE Iran (Fig.

2. 1; Alavi, 1994; Aghanabati, 2004; Alavi, 2004). This magmatic belt has been interpreted

as a subduction related Andean-type magmatic arc that has been active since the late

Jurassic within the collisional Alpine–Himalayan orogenic belt, reflecting subduction and

collision of the Afro-Arabian plate with Eurasia (Schroder, 1944; Dewey et al. 1973;

Dargahi et al., 2010). The rock units of this belt are composed of voluminous tholeiitic,

calc-alkaline, and K-rich alkaline intrusive and extrusive rocks, with associated pyroclastic

and volcanoclastic successions, formed along the active margin of the Iranian plate

(Berberian and King, 1981; Berberian et al., 1982). The belt hosts the largest of the Iranian

porphyry deposits, including Sarcheshmeh, Sungun, Meiduk, Dali, and Darehzar

(Shahabpour, 1994; Atapour and Aftabi, 2007; Boomeri et al., 2009).

The closure of the Neotethyan ocean and prevailing collisional tectonics during Tertiary

times built a highly fertile metallogenic environment with massive porphyry copper

deposits/prospects in the Urumieh–Dokhtar magmatic belt clustering in narrow arc

segments, typically a few tens of kilometres wide (e.g., Agard et al., 2005; McInnes et al.,

2005; Shafiei et al., 2009; Dargahi et al., 2010; Richards et al., 2012; Asadi et al., 2014).

Cenozoic tectono-magmatic activity and porphyry Cu-Mo mineralisation along the

Urumieh-Dokhtar magmatic belt are attributed to three time-windows: (1) Eocene–

Oligocene (Ahmadian et al., 2009); (2) mid-late Oligocene (Kirkham and Dunne, 2000;

51
McInnes et al., 2005); and (3) mid-late Miocene (McInnes et al., 2005; Razique et al.,

2007; Richards et al., 2012). The Urumieh–Dokhtar belt occurred during the Cenozoic

magmatism which started in late Cretaceous-Paleocene, peaked in Eocene and extended

into the Miocene and Quaternary. The magmatism was accompanied by the formation of

a wide range of ore deposits, consisting of epithermal ore deposits, skarn-type ores,

porphyry-type Cu-Mo-Au deposits and a variety of industrial minerals (Mirnejad et al.,

2010).

Most Iranian Cu porphyry deposits have been explored in the SE part of Iran especially

in Kerman province (e.g., Sarcheshmeh and Meiduk mines) and the NW part of Iran in

Azerbaijan province (e.g., Sungun deposit) since the 1970s. The central part of the

Urumieh-Dokhtar belt has recently received attentions for their porphyry-style ores. Few

porphyry Cu deposits are present in the central part of Urumieh–Dokhtar belt, typical

examples being Aliabad, Darehzereshk, Dali and Kahang (Zarasvandi et al., 2005; Ayati

et al., 2008).

2.2 Geology of the Kahang Deposit

The Kahang Cu-Mo porphyry deposit was initially discovered in 2003 from remote sensing

(Landsat TM) and geophysical studies and then from drilling (Tabatabaei and Asadi

Haroni, 2006; Afzal et al. 2012). Subsequently, stream sediment sampling, alteration

mapping and lithogeochemical exploration were undertaken as well as geophysical

exploration using induced polarization (IP) and resistivity (RS) which showed the

existence of a Cu-Mo prospect with Cu and Mo average grade of 0.1 wt.% and 33 ppm,

respectively (Afzal et al., 2010). This led into further subsurface exploration to find out if

52
there would be a deposit in this area. On the basis of alteration assemblages, the Kahang

prospect was divided into three divisions namely; Eastern, Central and Western Kahang

(Fig. 2. 4a). Within these, 48 boreholes were drilled in the Eastern Kahang with total depth

of about 22,000 m. There is a Cu resource greater than 100 Mt of sulphide ore with a Cu

mean value equal to 0.23 wt.% if the Cu threshold is 0.1 wt.% (See Chapter 3, Fig. 3. 18)

so the Kahang is not a prospect, and can appear to be promising.

The Kahang deposit lies within Eocene volcanic–pyroclastic rocks, which have been

intruded by Oligo-Miocene porphyritic granitoid rocks, quartz monzonites, monzodiorite-

monzogranites and diorites (e.g., Alavi, 1994; Tabatabaei and Asadi Haroni, 2006: Fig.

2. 1a and b). The Eocene rock units consist of andesite, trachyte, trachy-andesite, silicic

breccias and tuffs. The main geological structure in the area is a NW–SE and NE–SW

trending fault system. The extrusive rocks, including tuffs, breccias and lavas, are dacitic

to andesitic in composition. Magmatic events in the Kahang area have been interpreted

as followings (Afzal et al., 2010):

1. Explosive eruptions of pyroclastics such as tuffs and tuff breccias.

2. Flows of andesitic to dacitic lavas with porphyritic textures. It is probable that eruptions

of pyroclastic rocks and lavas were repeated periodically.

3. Emplacement of sub-volcanics and intrusive rocks with dacitic, andesitic, monzonitic

and dioritic compositions.

53
Fig. 2. 1. a) Geological map of the Kahang study area, scale: 1: 10,000 (Alavi, 1994; Tabatabaei
and Asadi Haroni, 2006), and b) structural map of Iran, showing the Urumieh-Dokhtar volcanic
belt (Alavi, 1994)

2.2.1 Mineralisation Characteristics of the Kahang Deposit

The Kahang deposit is a Cu-Mo porphyry deposit. Mineralisation is mainly hosted within

Eocene sub-volcanic rocks, especially porphyritic quartz diorites, monzodiorite-

monzogranite and dacitic rocks (Tabatabaei and Asadi Haroni, 2006; Afzal et al, 2011,

54
2012 and 2013b: Fig. 2. 2). Ore minerals are dominated by chalcopyrite, pyrite, bornite

and lesser amounts of chalcocite, covellite, malachite, molybdenite and Fe ores (i.e.,

hematite, magnetite, goethite and jarosite: Fig. 2. 3c).

Fig. 2. 2. Eocene sub-volcanic rocks in the Kahang deposit (View towards SE)

(a)

55
(b)

(c)

56
 (d)
Fig. 2. 3. a) Pyrite (Py), chalcopyrite (Ccp) and molybdenite (Mol), b) Copper secondary
sulphides and carbonate minerals from the oxidised zone (Azadi et al., 2014), c) Fe-oxides, and
d) reflected light photomicrograph showing bornite (Brn), chalcocite (Chl) and chalcopyrite
(Cpy) in the Kahang porphyry deposit

Based on vein morphology, mineral paragenesis and cross-cutting relationships, seven

groups of veins and veinlets were distinguished in the Kahang deposit (Afzal et al., 2010

and 2012; Azadi et al., 2014) namely; (1) Early biotite veinlets followed by (2) magnetite-

chlorite ± quartz ± sericite veins, (3) quartz-magnetite ± chlorite ± chalcopyrite ± pyrite

veins, (4) quartz-molybdenite ± chalcopyrite ± pyrite veins, (5) pinkish anhydrite-

chalcopyrite ± pyrite ± white anhydrite ± gypsum veins, (6) quartz-sericite-pyrite ±

chalcopyrite ± chlorite veins, (7) tourmaline ± quartz ± chalcopyrite ± pyrite veins and (8)

late poly-mineral calcite-sphalerite ± galena ± pyrite ± chalcopyrite veins (Afshooni et al.,

2010, 2011 and 2013; Azadi et al., 2014). The main mineralisation at the Kahang deposit

is Cu-Mo porphyry that occurs within intrusive bodies and their surrounding sub-volcanic

57
rocks. The ore minerals, consisting of chalcopyrite, pyrite, malachite, magnetite, limonite,

jarosite, goethite, bornite, sphalerite, galena, digenite, covellite, hematite, chalcocite and

molybdenite are distributed in leached, oxidised, supergene and hypogene zones (e.g.,

Berberian and King, 1981; Alavi, 1994; Ayati et al., 2008; Afshooni et al., 2010, 2011 and

2013; Asadi et al., 2015), as depicted in Fig. 2. 4. Gold occurs as fine inclusions within

pyrite and chalcopyrite and as native gold (electrum) within grey quartz veins in hypogene

zone (Fig. 2. 4L). Drilling data shows that a large-scale Cu–Mo mineralisation also occurs

in the hypogene zone. Pyrite in the hypogene zones generally occurs as aggregates,

composed of optically homogeneous euhedral to subhedral crystals, ranging in size from

20 μm to 5 mm which occurs in two generations: early pyrite (Py I) that is small rounded

blebs (~20–50 μm) included in chalcopyrite crystals (Fig. 2. 4a) and late pyrite (Py II)

distributed widely and formed later than chalcopyrite (Fig. 2. 4b). Chalcopyrite is the most

common sulphide mineral in the Kahang deposit, and appears as small rounded blebs

(50 μm–1 mm) as depicted in Fig. 2. 4b. Cu values increase within chalcopyrite especially

in the deeper parts of the deposit (See Chapter 3, Fig. 3. 27e).

58
 Fig. 2. 4. Photomicrographs of ore minerals in the Kahang deposit (Afshooni et al.,
2013): a) type II pyrite (Py II) associated with chalcopyrite, b) replacement of pyrite by
magnetite in type I pyrite (Py I), c) subhedral sphalerite, containing chalcopyrite
inclusions enclosed by pyrite, d) magnetite grains associated with Ti-mineral, pyrite and
chalcopyrite, e) ex-solution between chalcocite and chalcopyrite, f) hematite blades, g)
galena grains associated with chalcopyrite, h) covellite occurs as fracture-filling in pyrite,
i) malachite occurs as fracture-filling in micro-diorites, j) bornite together with
chalcopyrite, k) chalcocite, digenite and pyrite occur as veins, l) Backscattered electron
Image showing native gold (electrum) grains within late stage grey quartz

59
2.2.2 Hydrothermal Alteration

Hydrothermal alteration in the Kahang deposit (Western, Central and Eastern parts) is

pervasive, occurring in an area greater than 10 km2. Detailed alteration mapping shows

four major types of hydrothermal alteration: potassic, phyllic, argillic and propylitic (Figs.

2. 5a) based on detailed studies of the mineralogy and petrography of drill cores and

surface samples (Harati et al., 2013). As a result, hydrothermal alteration zones in the

Kahang deposit can be divided into four types (Afshooni et al., 2013; Azadi et al., 2014):

1) Early potassic alteration (K metasomatism) which occurs within and proximal to

mineralised veins and intrusions that contain Cu-Mo mineralisation, 2) Medial quartz-

sericite-pyrite (phyllic) alteration that partially overprints the early potassic alteration zone

and contains mineralised veins, 3) Argillic alteration in the outer and peripheral parts of

the altered and mineralised zone that overprints the previous alteration zones, 4)

Peripheral propylitic alteration of mainly sub-volcanic rocks, distal to the zone containing

mineralised veins and breccias. Cu-Mo-Fe sulphides are spatially and temporally

associated with the potassic and phyllic assemblages which include chalcopyrite,

molybdenite and pyrite (Afshooni et al., 2010, 2011 and 2013; Harati et al., 2013). The

alteration zones in this deposit follow the conceptual model of the alteration zones which

was proposed by Lowell and Gilbert (1970: Fig. 2. 5b).

60
 (a)

(b)

Fig. 2. 5. a) Alteration map of the Kahang deposit (Western, Central and Eastern parts; Harati et

al., 2013) and b) Conceptual model of Lowell and Gilbert (1970)

61
The potassic alteration zone is located in the central part of the deposit with neo-formed

biotite and KF veinlets (Fig. 2. 6a). This alteration and associated hypogene

mineralisation mainly occurred within the deepest and central parts of the zone containing

mineralised veins and breccias, within quartz diorite and quartz monzonite (Harati et al.,

2013; Azadi et al., 2014). The common mineral assemblage within the potassic zone

contains secondary biotite (S-Bt: Fig. 2. 6b), K-feldspar (Kf), quartz, sericite, pyrite,

chalcopyrite, bornite, magnetite and lesser amounts of anhydrite, chlorite, zircon, rutile

and hematite. Potassic alteration in this area is characterised by K-feldspar and irregularly

shaped crystals of Mg-rich biotite (secondary biotite) within volcano-plutonic rocks (VPR).

Petrographic observations and microprobe analyses point to the presence of two

compositionally distinguishable types of biotite within this alteration zone: 1) primary

biotite, which is Fe-enriched, brown in colour, and generally euhedral and 2) hydrothermal

biotite (Fig. 2. 6c), which is mainly pale-brown to greenish-brown in colour and very

ragged (Shahabpour, 1982). The hydrothermal biotite occurs interstitial to feldspar and

quartz and locally replaces amphibole and primary biotite phenocrysts (Khayrollahi,

2003). Replacement biotite was formed commonly by the alteration of amphiboles (Fig.

2. 6d).

62
 Fig. 2. 6. Photomicrographs of the potassic alteration zone of the Kahang deposit: a) neo-
formed biotite (Bt) and KF veinlets, b) secondary biotite (S-Bt) and quartz (Qtz), c) secondary
biotite–chlorite assemblage after igneous amphibole associated with the potassic alteration
(Afshooni et al., 2013), and d) fine-grained biotite as pseudomorphs of amphibole phenocryst,
and coarse-grained biotite cut by a quartz veinlet (Afshooni et al., 2013)

63
The phyllic alteration zone, developed in the eastern part of the deposit, contains high

amounts of quartz, sericite and albite within an argillic matrix. The phyllic alteration is

within acidic to intermediate sub-volcanic domes (Harati et al., 2013; Asadi et al., 2014).

The pervasive feldspar-destructive phyllic alteration is characterised by sericite, quartz,

pyrite, as main minerals and chlorite as an accessory phase. High abundances of quartz

in this zone are present within several generations of quartz stockwork veins, veinlets and

disseminations. Sericite also occurs as very fine grained to fine grained yellowish grains

within groundmass, veins and veinlets (Fig. 2. 7a). Chalcopyrite, zircon, rutile and some

clay minerals may be present. In sericitized rocks, K-feldspar is usually transformed into

sericite or fine-grained muscovite (Fig. 2. 7b) while biotite and amphibole are converted

to chlorite (Fig. 2. 7c). Phyllic alteration grades gradually into argillic alteration as

indicated by increasing amounts of clay minerals. Mineralogical changes of typical phyllic

alteration occur in granite–granodiorite in the study area.

64
 Fig. 2. 7. Photomicrographs of phyllic alteration zone in the Kahang porphyry deposit: a)
presence of sericite (Ser) and quartz (Qtz), b) plagioclase (Plg) phenocrysts are pervasively
replaced by sericite and surrounded by quartz grains in the phyllic alteration zone (Afshooni et
al., 2013), and c) biotite (Biot) altered to chlorite (Chl) in the phyllic alteration zone (Afshooni et
al., 2013)

65
The argillic alteration zone contains intermediate to high levels of alunite (especially in

the western and central parts of the deposit) indicative of silicified epithermal alteration

within the uppermost part of the deposit, predominantly in sub-volcanic rocks and

porphyry and dacite–rhyodacite stocks. This alteration zone is associated with the

formation of the clay minerals by extreme base leaching of alumino-silicate minerals (Fig.

2. 8a and b). This zone is represented by kaolinite, illite, and montmorillonite that replaced

plagioclase and mafic minerals in andesites and tuffs. Clay occurs as fine grained white

to brown coloured patches with increasing amounts of iron oxides within surface outcrops

and the outer parts of altered rocks. Jarosite is the second major alteration mineral in this

zone. In some places jarosite occupies a huge vol.% (≥ 50%) of the rock probably

indicating that it has undergone advanced argillic alteration (e.g., Azadi et al. 2014; Fig.

2. 8c).

66
Fig. 2. 8. Photomicrographs of argillic alteration zone in the Kahang porphyry deposit: a) argillic
alteration with clay minerals, altered plagioclase (Plag) and opaque minerals (Op), b) K-feldspar
(Kf) phenocrysts partially replaced by clay minerals (fine grey material) in argillic alteration zone
(Afshooni et al., 2013), and c) Advanced argillic alteration containing jarosite (Jar) overprinted
on quartz-sericite (phyllic) alteration (Azadi et al. 2014)

67
The propylitic alteration zone marks the outer limit of the hydrothermal system and is

dominated by chlorite, epidote and albite (Fig. 2. 9). This alteration zone is more

developed in margins and upper parts of the deposit. The propylitic alteration is

characterized by chlorite, epidote, albite, calcite, sericite, quartz, and clay mineral

assemblages that are locally replaced by biotite, plagioclase, hornblende and

groundmass.

Fig. 2. 9. Photomicrographs of propylitic alteration zone in the Kahang porphyry deposit:

a) plagioclase phenocrysts replaced by an aggregate of chlorite (Chl), epidote (Epi) and
calcite (Cal) in the propylitic alteration zone (Afshooni et al., 2013), and b) propylitic
alteration with pervasive epidote and chlorite

2.3 Dataset Particulars

From the outset, the 3D geological models for the Kahang deposit were created using

RockWorks™ v. 15 software with data from 48 boreholes. The data, manipulated in an

Excel database, included lithology, alteration type, and ore grades, were modelled using

the “Lithoblending” algorithm of the mentioned software. This subsurface data included

collar coordinates of drill cores, azimuth and dip (orientation), lithology, alteration,

mineralogy and zonation. The project dimensions were 600×660×780 m in the x, y and z

68
orientations and each voxel had a dimension of 4 m × 4 m × 10 m, respectively.

Topographical features of the deposit as well as other related data mentioned above were

formed into a 3D geological model (Fig. 2. 10).

Fig. 2. 10. The locations of drill cores with lithological units within the Kahang deposit and its 3D

surface topography

2.4 Assay Quality Assurance and Quality Control

Sampling is the fundamental part in a geochemical investigation for different stages of

mineral exploration and environmental purposes. The optimum sampling strategy, to

meet the company’s objectives, should be based on geochemical methods followed by

the field observations, variety of sampling, sample preparation and analytical approaches.

69
The estimate of reproducibility (precision) allows us to quantify variation of sampling and

laboratory analysis which is an integral part of the geochemical data interpretation. As a

result, any mistake in sampling and sample preparation may influence the results of the

survey (Thompson and Howarth, 1976; Fletcher, 1981; Demetriades, 2014).

From 48 drill holes in the Kahang deposit, 7146 lithogeochemical samples have been

collected at 2 m intervals. These samples were analysed using ICP-MS for 48 elements

by ALS Chemex (ALS Canada Ltd) and Zarazma Mineral Studies Company certified by

Geostats Pty Ltd (Australia: Appendix. A). Detection limits for Cu and Mo are 0.2 ppm

and 0.05 ppm, respectively. Moreover, 399 (Appendix. B) randomised samples for Cu

determination were selected and analysed for quality assurance and quality control

purposes, assessed using Thompson-Howarth error analysis (Thompson and Howarth,

1976 and 1978). The following procedure is suggested for estimation of precision from a

minimum of 50 pairs of duplicate samples (Thompson and Howarth, 1976):

(1) From the duplicate analyses, obtain a list of the means and absolute difference.

(2) Arrange a list (in Excel software) in increasing order of concentration means.

(3) From the first 11 results obtain the mean concentration and absolute difference of the

two results (controlling samples) from that group (each group contains 11

duplicated/reanalysed samples).

(4) Repeat step 3 for each successive group of 11 results, ignoring any remainder less

than 11.

70
(5) The mean of each replicate pair is plotted against the absolute difference between the

two analyses.

The highest value up the % scale on the right axis gives the precision. A precision around

5% is normal. If the precision is around 1%, the Y axis has not been properly calculated

with respect to the procedure mentioned above. The precision greater than 5% may have

cause for concern and reconsideration. However, the precision for Cu is around 2% in the

Kahang deposit with respect to 399 duplicated sample for Cu (Fig.2. 11; Appendix. B).

Fig. 2. 11. Estimation of precision of the Cu analyses using diagram of Thompson and
Howarth (1978). The mean of the replicate pairs is plotted along the X-axis, the absolute
difference of the two results along the Y-axis

71
2.4.1 Comparison of Geochemical Data Variances via F-Distribution
F-distribution test is used to identify variances equality of duplicated samples (e.g.,

geochemical data), which was introduce by the famous statistician, Sir Ronald Fisher

(1890–1962). This is the theoretical distribution of values which are expected by randomly

sampling from a normal population and calculating, for all possible pairs of sample

variances, the ratios as follow (Deutsch and Journel, 1998; Davis, 2002; Emery, 2012):

𝑆2
𝐹 = 𝑆12 S1 ≥ S 2
2

Equation 2-1

Where F, S1 and S2 represent F-distribution or continuous probability distribution and

variances for pair of samples (S1 = 0.222 and S2 = 0.219). The variances of double

samples vary if the number of observations used in their calculation is small. Therefore,

the shape of the F-Distribution is expected to change with changes in terms of samples

amounts.

The F-Distribution has two degrees of freedom equal to n1-1 and n2-1 in which n1 and n2

represent the number of observations equal to 398. Fisher showed that significance level,

1-α (α: probability value) is calculated in the cases of one-tailed and two-tailed

distributions depending on the defining alternative hypothesis. The hypotheses are as

follows (Fisher and Tippett, 1928; Emery, 2012):

72
 Null hypothesis: 𝐻0 : 𝜎12 = 𝜎22

Equation 2-2

Alternative hypothesis: 𝐻1 : 𝜎12 ≠ 𝜎22

Equation 2-3

Where σ1 and σ2 denote variances of populations. Based on the F-test, F(398,398) ≈

1.015 which is less than 1.2175 (obtained from Appendix. C) With respect to the

confidence level of 97.5% (α = 0.025). As a result, the Null hypothesis is acceptable

representing that two variances obtained from the paired samples are almost equal to

each other.

2.4.2 Comparison of Geochemical Data Means via Paired T-Test

A paired T-test is utilised to compare between means of two populations. The paired

sample T-tests typically include a sample of matched pairs of similar units (e.g., Cu wt.%

in this scenario), or one group of units that has been tested twice (e.g., Davis, 2002;

Emery, 2012; see Appendix. B).

The correct rejection of the null hypothesis (no difference between mean values) can

become much more likely. Because half of the sample now depends on the other half,

the paired version of Student's T-test has only "n/2–1" degrees of freedom (n is the total

number of observations). Pairs are individual test units and the sample has to be doubled

to achieve the same number of degrees of freedom.

73
To achieve the null hypothesis which the true mean difference is zero, the procedure is

as follows:

Calculate the difference between the two observations on each pair as follow:

di = yi − xi

Equation 2-4

2. Calculate the mean difference of the pair samples in terms of their grades (𝑑̅ ). The

grades means for the paired samples are 0.194% and 0.196% so 𝑑̅ is 0.002%.

3. Calculate the differences of standard deviation (Sd = 𝑆𝑑12 - 𝑆𝑑22 ) for the pair of samples.

To do this, the standard deviation of each sample (sd1 and sd2) was calculated and they

are equal to 0.468 and 0.472, respectively. Subsequently, standard error of the mean

difference was calculated (Equation 2. 5) which is 0.47.

𝑆𝑑
SE(𝑑̅ ) =
√𝑛̅

Equation 2-5

Where 𝑛̅ is equal to 2 because there is a pair of samples.

4. Calculate the T-test statistic under the null hypothesis, this statistic follows a T-

distribution with n − 1 degrees of freedom.

74
 𝑑̅
T= 2
SE(𝑑̅ )√𝑛
̅

Equation 2-6

Where 𝑛̅ is the number of paired samples which is 399 (See Appendix. B).

5. Use table of the T-distribution (Appendix. D) to compare value for T to the Tn−1

distribution. This will give a T critical (p-value), defined as the smallest level of significance

at which the null hypothesis would be rejected for a specific test, for the paired T-test

(Davis, 2002). The calculated T from paired samples is -0.06 according to the Equation

2. 6 and the T critical for “two-tailed test” resulted from Appendix. D with respect to

confidence level (probability value for α = 0.025) of 97.5% is equal to ∓1.9629 which

indicates that the Null hypothesis is again acceptable. Therefore, the mean values of the

paired samples are equal.

Consequently, results derived from T- and Fisher tests show that there is no significant

differences between results obtained via raw and controlling samples giving an analytical

accuracy in this deposit.

2.5 3D Geological Modelling of the Deposit

2.5.1 Lithological Model
Major rock types in the eastern part of the deposit are sub-volcanic units such as andesite,

andesite–diorite, dacite, diorite, granodiorite, quartz andesite, quartz andesite–diorite,

75
porphyric quartz diorite and tuff (See list of abbreviations). Dacitic rocks host ores in the

SE part of the study area.

The most heavily mineralised rocks are composed of porphyritic quartz diorite (Fig. 2. 12),

accompanied by granodiorite and dacitic rocks. The quartz-diorites are porphyritic

containing phenocrysts of plagioclase, biotite and rounded quartz.

The 3D lithological models in Figs. 2. 12a; 2. 13a; 2. 14a and c were generated using the

“Lithoblending” algorithm of RockWorks™ v. 15 software using data from 48 boreholes.

RockWorksTM uses a specific lithology modelling algorithm to do this extrapolation. As a

result, “lithoblending” is a solid modelling method that is utilised for generating geological

solid models (lithology, alteration, mineralisation and zonation) by the RockWorks TM

software which assigns the solid model by looking outward horizontally and vertically from

each borehole. The “lithoblending” first assigns the voxels immediately surrounding each

borehole according to the closest geological units (e.g., lithology). Then, it moves out by

a voxel to other neighbouring voxels located in one lithological unit or mineralised zone

and this action will continue in this manner until the program encounters a voxel that is

already assigned (Sweetkind and Drake, 2007; RockWorksTM 15, 2010; Amit et al., 2014;

Eslamian, 2014). Following this, the RockWorks™ v. 15 software looks at the observed

lithology intervals (2 m samples’ interval in the Kahang deposit), which are viewed in logs

and log sections already, and extrapolate the lithology throughout the project, outward

from the boreholes. This modelling process basically fills in the blanks between the logs.

76
The Borehole Manager Fence tools of RockWorksTM are available within the geological

model (e.g., lithology) which is utilised to display one or more vertical slices from the

inside of a lithological solid model. Subsequently, a lithological fence diagram was built

up using RockWorksTM software based on the constructed lithological model (Fig. 2. 12b).

This fence diagram contains two sections, showing NW-SE and SW-NE trends, which

contain high volume of PQD within the deposit. Furthermore, the andesitic rocks are

marginal in this area (Fig. 2. 12b).

(a)

77
 (b)
Fig. 2. 12. a) 3D lithology model of the Kahang eastern part (Yasrebi et al., 2012), and
b) lithological fence diagram (See abbreviation list for more details)

2.5.2 3D Alteration Model

Phyllic is the most spatially extensive alteration style in the eastern part of the Kahang

deposit, as illustrated in Fig. 2. 13. Potassic alteration is situated at depth and it is deeper

towards the eastern part of the deposit, although it does exist near surface in the western

part. Argillic and propylitic alterations are low in terms of their volume and occur near

surface as illustrated in Fig. 2. 13a. In addition, quartz-sericite and sericite alteration

zones are scattered throughout the deposit.

An alteration fence diagram was created based on alteration model, as shown in Fig. 2.

13b. The fence diagram includes two NW-SE and SW-NE sections which reveal that

78
phyllic alteration is within the uppermost part of the deposit. Potassic alteration is limited

and scattered throughout the deposit at depth. Moreover, argillic alteration is dominant

along the SE margin of the area.

(a)

79
 (b)
Fig. 2. 13. a) Alteration model of the Kahang eastern part (Yasrebi et al., 2012), and b)
alteration fence diagram (See abbreviation list for more details)

2.5.3 Ore-type Zonation Model

The main criteria for determining of mineralised zones is their index ore minerals. Index

ores consist of chalcopyrite and pyrite for hypogene, chalcocite, bornite and covellite for

supergene enrichment and malachite, azurite, tenorite and cuprite for oxidation zone

(Robb, 2005; Berger et al., 2008; Mihalasky et al., 2013). Studies of the pattern of zonation

in the eastern part of the Kahang deposit show that the most significant mineralisation (in

terms of ore zone size) is hypogene containing a high percentage of chalcopyrite

accompanied by pyrite. This can be easily seen in the 3D models in Fig. 2. 14c and d.

Molybdenite is present as vein and veinlet mineralisation with pyrite and chalcopyrite in

the hypogene zone (Fig. 2. 14f). The fence diagrams for mineralisation and zonation

80
models show the existence of high amounts of chalcopyrite and pyrite within the

hypogene zone (Figs. 14b and d).

(a)

(b)

81
(c)

(d)

82
 (e)

(f)
Fig. 2. 14. a) 3D ore-type zonation model, b) fence diagram of ore-type zonation model, c) 3D
dominant ore minerals, d); fence diagram of dominant ore minerals, e) chalcopyrite in stockwork
copper mineralisation from hypogene zone, and f) pyrite (Py) and molybdenite (Mol)
mineralisation in hypogene zone (See abbreviation list for more details)

83
CHAPTER THREE. Concentration-Volume (C-V)
Fractal Modelling for Separation of Mineralised
Zones

84
3.1 Introduction

Identification of supergene enrichment, hypogene, oxidized, leached zones from ‘barren’

host rocks (weakly mineralised zones) is one of the major purposes of ore deposit

modelling (Cheng et al., 1994; Li et al., 2003; Gałuszka, 2007; Makkonen et al., 2008;

Zeng et al., 2009; Afzal et al., 2012). Conventional geological methods for zone

recognition in porphyry deposits are generally based on mineralogical, petrographical and

alteration criteria (Schwartz, 1947; Beane, 1982; Sillitoe, 1997; Berger et al., 2008). A

conceptual model for lateral and vertical variations in alteration style was initially

suggested by Lowell (1968) and later by Lowell and Guilbert (1970), based on deposits

in the North American Cordillera Orogenic Belt. These models have been further

developed by Cox and Singer (1986) and Melfos et al. (2002). In addition, fluid inclusion

(e.g., Roedder, 1971; Nash, 1976; Ulrich et al., 2001; Asghari and Hezarkhani, 2008) and

S isotope studies (Wilson et al., 2007) have been utilised for determination of different

zones within porphyry Cu deposits. The mentioned models above do not consider the

distribution of elemental concentrations within ore deposits and do not rely on resource

modelling (See chapter 1, sections 1.1 and 1.3 for more information). It is a fact that ore

grades vary with changes in geological properties such as mineralogy and alteration

zones in porphyry Cu deposits (Zarasvandi et al., 2005; Berger et al., 2008; Pirajno, 2009;

Mihalasky et al., 2013; Xiao et al., 2014). Different geological clarifications would be

presented for defining boundaries of different zones in porphyry Cu deposits which may

also lead to different results if the ore element grade distribution is not taken into

consideration (Afzal et al., 2013c; Wang et al., 2013; Soltani et al., 2014; Sun and Liu,

2014; Yang et al., 2015).

85
3.2 Geometry of Natural Processes

The famous Greek mathematician Pythagoras of Samos, around 2500 years ago,

suggested that natural processes and behaviours are in accordance with mathematical

principles (Zhmud, 1989; Neimark, 2003; Hejazi, 2005; Neto, 2006; Rainer and Ruff,

2013). This hypothesis was not accepted by most scientists until the twentieth century.

Nowadays, models and theories based on mathematics are widely used by geoscientists

to better interpret natural processes.

Lobachevsky and Bolyai (around 1830) established non-Euclidian geometry in the 19th

century which showed that nature is not just in compliance with Euclidian geometry.

Benoît Mandelbrot and Gaston Julia (1959) invented a new geometry called “Fractal”.

However, in the 1960s, Mandelbrot started writing about self-similarity in papers such as

“How Long the Coast of Britain Is”. Mandelbrot reached the point where measuring of any

length with a large scale (e.g., continental and regional) is more time consuming when

measured by a small scale (e.g., local scale mapping). However, how long this

measurement takes is relevant to the various feature changes of the coastline which does

not follow the regular dimensions. On the other hand, fractals are not limited to Euclidian

geometric patterns, but can also describe processes such as structure and texture of

minerals and rocks in thin/polished sections (Mandelbrot, 1983; Das and Edgar, 2005;

Afzal et al., 2013c). In addition, this theory presented a geometry in which the features do

not follow up dimensions with integer numbers and are not discrete, but they meet

decimal, irregular and continuous numbers. The general agreement is that theoretical

fractals are infinitely self-similar, iterated and detailed mathematical features containing

86
fractal dimensions, of which many examples have been formulated and examined

precisely (Mandelbrot, 1983; Falconer, 1991; Falconer and Hu 2001). The self-similarity

of a feature is its shape on a much smaller scale, as depicted in Fig. 3. 1. Fractal patterns

with different degrees of self-similarity have been rendered or studied in images, sounds

and structures in nature (Turcotte, 1986; Cheng et al., 1994; Cheng, 1999; Zuo et al.,

2009a; Deng et al., 2010; Wang et al., 2010a and b; Afzal et al., 2011and 2014; Yasrebi

et al., 2013a, b and 2014).

Fig. 3. 1. Self-similarity in a triangle

The word “fractal” was coined by Mandelbrot (1976) from the Latin word “fractus” or

“fractum”, meaning broken, which he has applied to objects that were too irregular to be

defined by ordinary Euclidean geometry (Davis, 2002). Mandelbrot (1983) wrote a famous

book entitled “Fractal Geometry of Nature” and introduced “fractal” as a new branch of

non-Euclidian geometry.

87
The fractal geometry of each shape and its complications are shown in the form of real

numbers, as in Euclidean geometry concepts of measured angle, length, area and

volume. There are fractal dimensions which are not integers and can be real and decimal

such as 1.4, 2.3 and 3.5 (Fig. 3. 2).

Fig. 3. 2. Changes in dimensions of a fractal shape of Koch Curve (Zhu et al., 2003)

Many studies have indicated that hydrothermal ore deposits such as porphyry Cu,

orogenic gold and epithermal polymetallic deposits present non-Euclidean variations in

ore element values in rocks, alterations and related surface materials such as water, soils,

stream sediments, humus and vegetation (Cheng, 2007; Cheng and Agterberg, 2009;

Afzal et al., 2011 and 2013a, c; Heidari et al., 2013; Yasrebi et al., 2013a; Soltani et al.,

2014). As a result, variation of fractal dimensions in geochemical data can furnish

complementary information and applicable criteria to delineate mineralised and ‘barren’

zones from host rocks within a studied area. Different log–log plots in fractal/multifractal

models are proper tools for separation of geological populations based on geochemical

88
data since threshold values (breakpoints) can be determined in those log-log plots (Cheng

et al., 1994; Agterberg et al., 1996). These geochemical threshold values are identified

via fractal analysis which is usually correlated using geological field observations (e.g.,

mineralisation, alteration, lithological units and ore seams). In other words, fractal analysis

is able to indicate differences within mineralisation, alteration, lithology and zonation of

ore deposits especially in hydrothermal occurrences such as porphyry Cu deposits

(Goncalves et al., 2001; Cheng, 2007; Carranza, 2008; Carranza et al., 2009; Cheng and

Agterberg, 2009; Afzal et al., 2011 and 2012). However, proper knowledge of the

geological and geochemical aspects of a deposit is important in order to identify

characteristics of geochemical populations on the basis of fractal analysis (Cheng, 1999;

Sim et al., 1999; Li et al., 2003; Carranza, 2009; Carranza and Sadeghi, 2010).

The aim of this chapter is to use a Concentration-Volume (C-V) fractal model to delineate

Cu and Mo mineralised zones in the Kahang porphyry deposit of Central Iran in order to

generate an optimised block model for determination of an ultimate pit limit and a best

mining strategy. Moreover, to correlate and validate the results, the outcomes of the

fractal models will be compared with geological models using a logratio matrix proposed

by Carranza (2011).

3.3 Statistical Characteristics

In the studied deposit, 7146 core samples were collected from 48 boreholes at 2 m

intervals, and analysed by inductively coupled plasma mass spectrometry (ICP-MS) for

Cu and Mo (See section 2.3 and Appendix. A). The Cu and Mo distribution functions are

89
not normal, with Cu and Mo averages of 0.166 wt.% and 28 ppm, respectively (Fig. 3. 3).

The elemental distributions show an L shape with most of the volume of the deposit

containing low grades for Cu and Mo. Most values of Cu and Mo are lower than 1 wt.%

and 200 ppm, respectively. Variation between maximum and minimum of these data

shows a wide range among elemental concentrations (Table. 3. 1). Based on the

abnormal elemental distributions, Cu and Mo medians are assumed to be equal to

threshold values for separation of ‘barren’ host rocks and mineralisation which are 0.08

wt.% for Cu and 9.9 ppm for Mo (Davis, 2002). In this deposit, 33 and 14979 samples

were determined from 11 and 42 boreholes respectively out of a total of 48 boreholes

carried out in the deposit for density and RQD analysis, respectively (Tables. 3. 2 and 3.

3). Figures for the original data sets used: Cu and Mo grades, RQD and density values

have been generated using MATLAB software, as depicted in Fig. 3. 4. Since the Kahang

deposit is at a pre-feasibility stage, and the main target is Cu, only 399 randomised

samples for Cu (section 2.4 and Appendix. B) were analysed. The error for Cu is 4.04%,

calculated from the following equation (Govett, 1983):

2 |𝑋𝑖−𝑌𝑖|
e = 𝑛 ∑𝑛𝑖=1 𝑋𝑖+𝑌𝑖

Equation 3-1

Where e and n are error value, amount of re-analysed samples (399 duplicated sample

for Cu). Xi and Yi denote measured values for duplicated samples (See Appendix. B,

second and fourth column for Cu wt.% in Table. B. 1 for Xi and Yi).

90
 (a)

(b)
Fig. 3. 3. Histograms for data from the Kahang deposit: a) Cu wt.%, and b) Mo ppm

Table. 3. 1. Statistical characteristics for Cu and Mo

Elements Minimum Maximum Standard

Range Mean Median Variance Skewness Kurtosis
Grades Value Value Deviation

Cu (wt.%) 0.0003 4.92 4.91 0.16 0.271 0.087 0.073 6.6 74.5

Mo (ppm) 0.5 1,479 1,478.5 28.27 76.178 9.9 5,803.132 8.4 96.1

91
Table. 3. 2. Density analysis from 11 boreholes in the Kahang deposit (See abbreviation list for
rock type)

Depth Rock Density Depth Density

BHID BHID Rock Type
(m) Type (t/m3) (m) (t/m3)

KAG_27 151.7 PQD 2.67 KAG_50 148.6 PQD 2.68

KAG_27 459.55 PQD 2.8 KAG_50 266.1 QAN-D 2.67

KAG_27 580.3 ANS 2.7 KAG_50 332.95 DAC 2.74

KAG_28 65.7 PQD 2.67 KAG_51 78.6 DAC 2.63

KAG_28 232.1 PQD 2.66 KAG_51 174.2 ANS-D 2.59

KAG_30 141.5 PQD 2.7 KAG_51 547.3 PQD 2.81

KAG_30 240.7 PQD 2.72 KAG_52 235.85 PQD 2.72

KAG_33 51.2 DAC 2.7 KAG_52 376.45 PQD 2.69

KAG_33 128.15 DAC 2.63 KAG_52 462.55 PQD 2.69

KAG_33 223.6 DAC 2.71 KAG_52 530 PQD 2.73

KAG_33 348.8 DAC 2.64 KAG_54 644.5 DAC 2.7

KAG_36 72.9 PQD 2.34 KAG_55 73.15 QAN 2.62

KAG_36 327.9 DAC 2.69 KAG_55 108.4 QAN 2.69

KAG_36 421.15 ANS 2.68 KAG_55 267.55 QAN 2.77

KAG_49 359.75 ANS 2.71 KAG_55 470.85 PQD 2.62

KAG_49 545.25 PQD 2.76 KAG_55 361.75 ANS 2.71

KAG-55 634.55 PQD 2.7

92
 Table. 3. 3. Distribution number of RQD samples among drill cores in the Kahang deposit

93
(a)

(b)

94
 (c)

(d)
Fig. 3. 4. 3D maps for original datasets: a) Cu wt.%, b) Mo ppm, c) density t/m3, and d) RQD %

95
3.4 Block Modelling

Choosing a suitable voxel size for evaluation of a reserve/resource is crucial for

minimising errors (Asghari and Madani Esfahani, 2013; Shahbeik et al., 2014). This

problem has been assessed for estimated block models using different geostatistical

methods such as ordinary kriging (OK) and inverse distance weighted (IDW). Results

obtained by the estimation methods relate to the determination of voxel size in block

modelling (David, 1970; Cressie, 1993; Soltani Mohammadi et al., 2012).

Utilising a larger voxel size will increase the averaging effect in the estimated block model

in terms of concentrations, geophysical data, rock mechanical data and other attributes.

Additionally, a smaller voxel size will show more details, but potentially more error in an

anisotropic environment (Journel and Huijbregts, 1978). On the other hand, reducing the

voxel size results in an increase in estimated errors (variance and standard deviation) for

the final block model. Moreover, increasing the voxel size in the block model changes the

higher or lower grades of mineralised zones by smoothing of these points with high or low

values within a large voxel.

Identification of an optimised voxel size is one of the most important aspects of building

an estimated 3D block model. Therefore, it is necessary to select an optimal voxel size

with respect to the deposit geometry and drilling pattern because most of the

geostatistical software, e.g. RockWorksTM which was employed in this study, estimates

an ultimate block model based on the closest points considering particular parameters

such as ore element concentrations (Verly, 1984; Journel, 1993).

96
Statistical parameters such as mean and median can be used for recognition of optimum

voxel dimensions in various types of ore deposits (David, 1970; Journel and Huijbregts,

1978).

David (1970) proposed an applicable method for an operation based on geometrical

particulars of the different types of ore deposits and grid drilling. Based on the method,

voxel dimensions are calculated as follows:

a) Length and width of each voxel is equal to between half and quarter of the distance

between the drill cores according to along the least variability deposit.

b) Height of each voxel is delineated due to the type of the deposit. In ‘massive’

deposits such as magmatic deposits (e.g., porphyry deposits), the parameter is equal to

the height of excavating benches in the open pit mines (Hustrulid and Kuchta, 2006).

For recognition of the optimum voxel dimensions in the Kahang Cu porphyry deposit,

statistical characteristics consisting of mean, median and median absolute deviation

(MAD) were utilised. In addition, standard deviation (SD) was used for further comparison

and validation through the obtained results in the different scenarios of voxel size. If SDs

include very low changes then voxel size selection is carried out based on the Cu

estimated mean. The voxel sizes with the lowest value of Cu mean should be selected

because this is a worse scenario for mine planning and exploitation due to conservative

97
mine strategy and risk analysis. Moreover, the median and MAD are used for

determination of voxel sizes and development of conventional method, as proposed by

David (1970).

The 2D map which indicated the location of 48 boreholes drilled in the Kahang deposit

was constructed by RockWorks™ v. 15 software (Fig. 3. 5a). From this, a grid model of

the boreholes on the surface was created to illustrate drill core locations including the

location information, symbol style and borehole names for the studied area. Since the

grid drilling pattern within this deposit is not homogeneous and systematic, 14 pairs of

closest boreholes were selected for an optimum voxel size investigation because this

action can improve the interpolation of voxel values (Cu grades in this scenario) that lie

between data point clusters (Fig. 3. 6). The particulars of these pairs are revealed in the

Table. 3. 4. The distance range of the selected boreholes varies between 5 to 27 m. For

identification of an optimum voxel size in the directions of X and Y, the vector analysis

was employed. The ranges of distances in the X and Y directions are 0.38-18.97 m and

3.47-25.97 m, respectively (Table. 3. 4). Based on the David (1970) method (as described

in the last page), the voxel size in the Z direction was determined as 10 m on the basis of

the ore deposit geometry and particularly height of the working bench.

98
(a)

99
 (b)

(c)

Fig. 3. 5. a) 2D, b) 3D, and c) Google Earth maps of the grid drilling in the Kahang deposit

100
In the first step, the mean of distances between selected borehole pairs were calculated,

as depicted in Table. 3. 5. In the simple method, the range of voxel sizes in the X and Y

directions was considered between half and quarter of the mean value which is equal to

4.67-9.34 m. Moreover, the ranges of the voxel size according to median and MAD values

are 5-10 m and 2.5-5 m, respectively.

Fig. 3. 6. Boreholes location (ignoring the three isolated boreholes in the NW) map in the

Kahang deposit with selected closest borehole pairs

101
In the second step, the voxel size was investigated via vector analysis between boreholes

based on the closest surface location on the 2D map. The mean values in the directions

of X and Y are 12.91 m and 11.76 m, respectively, meaning that the voxel size varies

between 3.23 m and 6.46 m in X and between 2.94 m and 5.88 m in Y. Based on the

median values, the voxel size value ranges are 3.8-7.6 m and 2.32-4.65 m in terms of X

and Y. The MAD values for X and Y are less than 3 m indicating that the voxel size is less

than 1 m, resulting in an increase in the error for the construction of a final block model

(Goovaerts, 1997). For a massive ore body, and homogenous distribution of element

concentrations in porphyry deposits, X and Y directions have equal values in terms of

voxel size (Davis, 2002). As a result, five different voxel size scenarios of 5 m x 5 m x 5

m, 4 m × 4 m × 10 m, 5 m × 5 m × 10 m, 10 m x 10 m x 10 m and 15 m x 15 m x 15 m

have been allocated to build the Kahang deposit pre-Cu block model. In order to find an

accurate voxel size, declustering should be conducted previously because it is believed

that the proper voxel size with respect to the different voxel alternatives is the one with

the minimum standard deviation. In addition, Cu mean of a pre-Cu block model with an

accurate voxels size should be close to Cu mean value obtained from the declustered

data (e.g., Fig. 3. 7; Table. 3. 6; Deutsch and Journel, 1998; Richmond, 2002; Emery and

Ortiz, 2005 and 2011; Olea, 2007; Sadeghi et al., 2015).

3.4.1 Cell Declustering

Data are often spatially clustered which makes it difficult to determine whether they are

representative of the entire area of interest (Fig. 3. 7). To obtain a representative

distribution, one approach is to assign declustering weights whereby values in cells with

102
more data receive less weight than those in sparsely sampled areas. The grid drilling in

the area is non-uniform and the data need to be declustered. This operation was carried

out using the Declus program which incorporates the GSLIB library (Deutsch and Journel,

1998). The Cu mean and standard deviation of the declustered data are 0.145 wt.% and

0.22077. Subsequently, the Cu block models were generated by IDW utilising

RockWorksTM software. For determination of optimum voxel dimensions based on the

statistical parameters depicted in Table. 3. 5, standard deviation (SD) and an average Cu

value have been calculated.

Fig. 3. 7. Cu histogram based on declustered data

103
 Table. 3. 4. The particulars of the selected borehole pairs in the Kahang deposit

Borehole ID Distance Distance (m)

From To (m) X Y

KAG50 KAG6 11 4.4 10

KAG6 KAG47 23 17.14 15.33

KAG15 KH-DDH17 5 0.38 4.98

KAG33 KH-DDH13 20 18.69 7.1

KAG42 KH-DDH9 14 12.88 5.47

KH-DDH14 KH-DDH15 21 16.32 13.21

KAG41 KH-DDH15 15 14.59 3.47

KAG59 KAG27 15 12.28 8.6

KAG54 KH-DDH16 16 15.58 3.59

KAG52 KAG18 28 17.77 21.63

KH-DDH11 KAG19 20 18.97 6.31

KAG51 KH-DDH02 22 7.41 20.71

KAG33 KH-DDH9 25 17.04 18.28

KAG48 KAG28 27 7.39 25.96

Table. 3. 5. Variation of voxel size based on mean, median and MAD

Statistical Parameters Total Distances (m) Distances in X (m) Distances in Y (m)

Mean 4.67-9.34 3.23-6.46 2.94-5.88
Median 5-10 3.8-7.6 2.94-5.88
MAD 2.5-5 <1 <1

104
Topographical features of the deposit were formed into a block model. The block model

for Cu was produced by applying an upper and lower filter using RockWorksTM software,

based on the surface data (Fig. 3. 8) and borehole data collar heights, given by NICICO.

Those voxels located above the upper and below the lower topographical surfaces and

bed rock are considered as the waste voxels and are not included in the deposit block

model as the voxels are of negative significance (Todorov et al., 2002; Popov et al., 2003;

Hustrulid and Kochta, 2006; Yasrebi et al., 2011). The use of IDW to construct the block

model was employed in this research. The amounts of voxels with positive values (Non-

Zero) are shown in Table. 3. 6. The more Non-Zero voxels consequently correspond to

the voxel dimension of 5 m × 5 m × 5 m. The standard deviation value for the voxel size

of 4 m x 4 m x 10 m is lower than other voxel alternatives (Table. 3. 6). Moreover, the

averages for estimated Cu values were compared and the lowest value occurs in the 4 m

× 4 m × 10 m block model which is conservatively suited for identification of Net Present

Value (NPV) and subsequently mine planning (Hustrulid and Kochta, 2006). The Cu mean

for 4 m × 4 m × 10 m block model (0.15823 wt.%) is relatively close to the Cu average

obtained from the declustered data (0.145 wt.%) in comparison with the other voxel

scenarios. (Fig. 3. 7 and Table. 3. 6).

105
Table. 3. 6. Voxel numbers and Standard deviations and averages of Cu for different block
models

Block Model
Total Voxel Non-Zero Voxel Standard Cu Average
Dimensions
No. No. Deviation (%) (wt.%)
(m3)

4 × 4 × 10 1,113,742 263,414 0.20134 0.15823

5 × 5 × 10 718,505 169,091 0.20136 0.15833

5×5×5 2,526,601 345,578 0.20352 0.15895

10 × 10 × 10 322,873 42,284 0.20138 0.15828

15 × 15 × 15 97,785 12,486 0.21641 0.16266

Fig. 3. 8. 2D topographical surface of the Kahang deposit

106
Borehole collar heights were compared with topographic surface heights to check that

they are comparable, as depicted in Fig. 3. 9. The coordinates of collar boreholes

correlate with used topographical points to provide a topographic surface for resource

modelling and further optimisation studies. For the optimised pit scheduling software, a

topography model is a three-dimensional surface model (See chapter 5, section 5.4)

which is analogous to wireframe surfaces in most mining software systems (NPV –

Scheduler, 2001).

107
Fig. 3. 9. Correlation between borehole data collar heights and topographical surface

108
3.5 Geostatistical Modelling

Geostatistical estimation methods are used commonly for interpolation and estimation of

different regional variables in 1D, 2D or 3D environments. Employment of an accurate

estimation method with respect to geometry and geological properties of different ore

deposits and also drilling patterns is a problematic issue in resource/reserve estimation

(David, 1970; Yasrebi et al., 2009; Shahbeik et al., 2014). Determination of estimation

methods is essential for decreasing the error estimation and increasing the accuracy of

resource and reserve evaluation (Dimitrakopoulos et al., 2007; Parhizkar et al., 2011).

Selection of an estimation method is essential for fractal/multifractal modelling, especially

in the C-V model. On the other hand, accuracy of the estimation methods and their errors

of interpolation affect the C-V fractal/multifractal modelling (Agterberg et al., 1993; Cheng

and Agterberg, 1996; Lima et al., 2003; Agterberg, 2012; Afzal et al., 2013a; Heidari et

al., 2013).

Linear and non-linear Kriging methods, Inverse Distance Weighted (IDW), have been

used to interpolate polynomials and splines, to overcome the mentioned problem (Franke,

1982; Zimmerman et al., 1999; Juan et al., 2011). Among these methods, kriging and

IDW are usually utilised in most cases in mineral exploration and mining engineering.

Using real data rather than synthetic data has several advantages; for example, it

precludes one method from having an unfair advantage merely because the data used

for the comparison is generated under the same model on which the method is based.

On the other hand, only with synthetic data can the effect of certain data characteristics

on interpolation accuracy be systematically evaluated (Englund et al., 1992; Weber and

109
Englund, 1992 and 1994; Zimmerman et al., 1999). Evaluation of ore element distribution

is an important parameter for mine planning and design (Hustrulid and Kochta, 2006).

3.5.1 Inverse Distance Weighted Anisotropic Method (IDWAM)

Inverse Distance (ID) is one of the more common gridding and estimation methods. With

this method, the value assigned to a voxel is a weighted average of either all of the data

points or a number of directionally distributed neighbours. The value of each of the data

points is weighted according to the inverse of its distance from the voxel (Zimmerman et

al., 1999; Homayoon et al., 2010; Shahbeik et al., 2014).

Inverse Distance Weighted Anisotropic (IDWAM) is a method for interpolation of scattered

points that estimates voxel values (e.g., ore grade) by averaging the values of sample

data points in the neighbourhood of each processing voxel. IDWA has a crucial

assumption that the interpolating surface is mostly influenced by the nearby points and

less by the more distant points. The interpolating surface is a weighted average of the

scatter points and the weight assigned to each scatter point diminishes as the distance

from the interpolated point to the scattered point increases. The main advantage of the

IDW method is to produce a smooth and continuous grid and does not exaggerate

extrapolations beyond the given data points (Franke, 1982; Goovaerts, 1997).

Therefore, the IDWAM is recommended for geochemical mapping where the data

boundaries (geochemical populations) are critical for threshold-based target separation

(e.g., in C-V fractal models). The range of element concentrations will be smaller than the

raw data range meaning that highest grade values will be less than the maximum of raw

110
data, and the lowest grade values will be greater than the minimum data point (Tahmasebi

and Hezarkhani, 2010). A general form of finding an interpolated value at a given point

based on samples for using IDW is an interpolating function:

Equation 3-2

where

Equation 3-3

x denotes an interpolated (arbitrary) point, xi is an interpolating (known) point, is a given

distance (metric operator) from the known point xi to the unknown point x, N is the total

number of known points used in interpolation and is a positive real number, called the

power parameter (e.g., an exponent of “2” = Inverse Distance Squared, “3” = Inverse

Distance Cubed; Shepard, 1968). The greater the value of the exponent, the less

influence distant control points will have on the assignment of the voxel value.

111
The disadvantages of conventional IDW methods are choice of weighting function which

may introduce ambiguity; especially where a fixed search radius requires a

neighbourhood distance and a minimum or maximum number of points.

In the IDWAM, all points will be used which increases error in the form of under and over

estimation so the samples located in a supergene enrichment zone (in a porphyry deposit)

can be influenced by the leached zone which correspondingly reduces the voxels grade

values within ore minerals consisting of say chalcocite, covellite and bornite (Pirajno,

2009). To overcome this problematic issue, variography, in combination with IDWAM, is

employed for better estimations.

Directional and non-directional searching in this method can improve the interpolation of

voxel values that lie between data point clusters and be useful for modelling drill-hole

based data in the stratiform and massive ore deposits (Zimmerman et al., 1999). In this

thesis, a combination of IDWAM and variography has been used in order to generate a

block model in terms of Cu and Mo values based on the following criteria:

1- The grid drilling pattern is irregular and non-systematic, with an especially high drilling

density in the NE part of the deposit, and low density in the NW part (e.g., three isolated

boreholes: Fig. 3. 5a, b and c). Moreover, the grid drilling pattern has an anisotropic

geometrical shape.

112
2- There are too many scattered drill holes in the marginal parts of the deposit which leads

to a lack of data.

3- Simple Kriging (SK), as a common estimation method, is based on a moving average

of the variable of interest, appropriate for various dispersed forms of data e.g. sparse

sampling points. As a result, this estimator requires adequate drill holes and data analysis

which are not met in the Kahang deposit. However, in this thesis a combination of IDWAM

and variography (horizontal and vertical) has been carried out for the development of

IDW.

4- According to field observations, the mineralisation and alteration zones, particularly

hypogene and phyllic, continue through to the marginal parts of the deposit (especially in

the SE area with high Cu and Mo values) which were not covered by the 48 boreholes

conducted in the studied area. The IDWAM is therefore appropriate due to the lack of

subsurface information.

5- Trends of Cu values in X, Y and Z show that there is no association between ore grade

and X-Y location or depth within the deposit (Fig. 3. 10), indicating again that “Universal

Kriging” is not appropriate for this deposit.

113
(a)

(b)

114
 (c)

Fig. 3. 10. Scatterplots for correlation between Cu (wt.%) and coordinates: a) Cu values
trend in X, b) Cu values trend in Y, and c) Cu values trend in Z

3.5.2 Application of IDWAM

The experimental variograms in Fig. 3. 11 in horizontal (Azimuth: 0 and Dip: 0) and vertical

(Azimuth: 0 and Dip: -90) directions were generated using MATLAB software with respect

to log transformations of Cu and Mo grades. The horizontal and vertical ranges for Cu are

56 m and 270 m, respectively. Moreover, the Mo horizontal and vertical ranges are equal

to 40 m and 80 m, respectively. The spherical model was fitted to the experimental

variograms. Accordingly, the theoretical variograms for Cu and Mo grade values are as:

115
 𝛾𝐶𝑢 (ℎ) = 0.41 + 0.36 𝑠𝑝ℎ (10, 10, 25) + 0.85 𝑠𝑝ℎ (56, 56, 270)

Equation 3-4

𝛾𝑀𝑜 (ℎ) = 1.125 + 0.56 𝑠𝑝ℎ (10, 10,10) + 0.36 𝑠𝑝ℎ (40,40, 20) + 0.044 𝑠𝑝ℎ (∞, ∞, 80)

Equation 3-5

(a)

116
 (b)

Fig 3. 11. Experimental and theoretical variograms: a) Cu, and b) Mo

In addition, ‘Variogram Maps’ have been created using Varmap software from the GSlib

Library (Deutsch and Journel, 1998) in order to find out if there are any anisotropic

characteristics within the deposit. Variograms are traditionally constructed as 1D curves:

ɣ (h) as a function of the distance h along a particular direction. It is often useful to have

a global view of the variogram values in all directions (X-Y, Y-Z and X-Z in this thesis). To

do this, transformed Cu and Mo grades were calculated using a logarithmic function

before variography was carried out in the cases of: one set for the X-Y orientation (with

lag spacing of 15 m) to evaluate the range in plan view, and one set either X-Z and Y-Z

with lag spacing of 8 m to evaluate the range in vertical sections. There is an isotropic

behaviour in horizontal direction for Cu and Mo. Furthermore, there is a weakly

117
anisotropic behaviour in the studied area based on the X-Z and Y-Z (vertical) variogram

maps, as shown in Fig 3. 12.

(a)

(b)

118
(c)

(d)

119
 (e)

(f)
Fig. 3. 12. Variogram maps for a) Cu in plan 1730 m, b) Mo in plan 1730 m, c) Cu in E-W
section with Northing = 3644585, d) Mo in E-W section with Northing = 3644585, e) Cu in N-S
section with Easting = 638325, and f) Mo in N-S section with Easting = 638325

120
The Kahang deposit was modelled with 489,927 voxels with each voxel having a

dimension of 4 m × 4 m × 10 m in the X, Y and Z directions. The project dimensions are

600, 660 and 780 m (Fig. 3. 13). 3D block models for Cu and Mo were evaluated by

IDWAW using the RockWorksTM software package, as depicted Fig. 3. 15. The ranges of

Cu and Mo from the variograms in Fig. 3. 11 were imported into the vertical and horizontal

distance cut-offs in the “Solid Modelling Options”.

In order to achieve this using the RockWorksTM software, the following tasks, in their

relative order, were carried out (Fig. 3. 14):

The Weighting Exponent value was determined as being equal to 2 in order to prompt to

enter a real number value for the Inverse-Distance exponent. Number of neighbouring

points were defined between 3 and 15 data points that were to be used when computing

the voxel value. The horizontal and vertical ranges (known as ‘Cut-offs Distances’ in the

RockWorksTM software) were recognised based on the combined variograms with lags’

spacing of 15 m and 8 m for horizontal and vertical directions (Fig. 3. 11).

121
Fig. 3. 13. Project dimensions of the studied area

Fig. 3. 14. Steps of IDWAM run in RockWorksTM 15

122
The sections within 0.4 wt.% Cu and 200 ppm Mo were separated and showed that these

parts were located in the central, NE and NW parts of the area (Fig. 3. 16). Additionally,

Cu values higher than 0.4 wt.% exist in the SE part of the area and continue towards

Kahang village.

Histogram and statistical characteristics of the Cu estimation indicate a standard deviation

of 0.15 which is lower than the IDW (See Table. 3. 6), as shown in Fig. 3. 17 and Table.

3. 7. As a result, it shows that the variography decreases the values of error estimation in

the IDW indicator. Histogram and statistical particulars of the Mo estimated model (Fig.

3. 17) illustrate a lower standard deviation in comparison with raw data which are 0.56

and 0.76 respectively. Moreover, the mean of the Cu and Mo in the estimated models are

0.14 wt.% and 27.49 ppm respectively (Table. 3. 7). If the Cu threshold is equal to 0.1

wt.% then the resource is greater than 100 Mt with a Cu mean value equal to 0.23 wt.%,

as depicted in Grade-Tonnage (G-T) diagram (Fig. 3. 18). The G-T diagram shows that

the deposit has a good potential as a porphyry type because the mean Cu in the different

thresholds are comparable with other Iranian porphyry deposits such as Masjed Daghi,

Darreh Zar and Sar Kuh with 340 Mt and 0.27 wt.% Cu, 475 Mt and 0.36 wt.% Cu and

110 Mt and 0.26 wt.% Cu, respectively (e.g., Shahabpour, 1994; Afzal et al., 2011;

Aghazadeh et al., 2015). Several plans were generated for Cu and Mo, as depicted in

Fig. 3. 19. The plans reveal that the main mineralisation occurs in the SE, NE and central

parts of the deposit, especially in the NW section of the studied area (area around the

three isolated boreholes: Fig. 3. 20). Cu values higher than 0.25 wt.% (Fig. 3. 20c) in the

isolated bore holes (NW part) are more common at depth, and it may therefore be wise

123
to drill more bore holes in this area. Moreover, Cu samples with values higher than 0.15

wt.% are mostly accumulated in the central part of the deposit (Fig. 3. 20b). Cu values

greater than 0.07 wt.% are scattered in majority parts of the deposit (Fig. 3. 20a).

Table. 3. 7. Statistical characteristics for Cu and Mo estimated models

Minimum Maximum Standard

Element Range Mean Median Variance Skewness Kurtosis
Value Value Deviation

Cu
0.0003 4.00773 4 0.14 0.151 0.091 0.023 4.38 39.60
(wt.%)

Mo
0.65 790.82 784.75 27.49 56.44 12.56 3,185.515 6.82 59.91
(ppm)

(a) (b)
Fig. 3. 15. Block models in the Kahang deposit: a) Cu, and b) Mo

124
 (a) (b)
Fig. 3. 16. Distribution models in the Kahang deposit: a) Cu ≥ 0.4 wt.%, and b) Mo ≥ 200 ppm

Fig. 3. 17. Histograms of estimated element concentrations: a) Cu, and b) Mo

125
Fig. 3. 18. a) Cu (%) Grade-Tonnage, and b) Cu (%) average-cut-off diagrams for Kahang
deposit

126
127
128
129
130
Fig. 3. 19. Cu and Mo plans in the Kahang deposit

(a)

131
 (b)

(c)

Fig. 3. 20. Samples (black discs) within Cu values higher than: a) 0.07 wt.%, b) 0.15 wt.% and
c) 0.25 wt.% from boreholes carried out in the deposit

132
3.6 Validation Processes

For validation of the estimation, two methods were used: optical correlation and jackknife.

Optical correlation is a visual verification which tries to visually show correlation coefficient

between raw (original) and estimated data in cross-sections (Fig. 3. 21). It shows a

positive relationship between Cu values from the boreholes (raw or original data) and the

generated block model (estimated data) in the different cross-sections (Fig. 3. 21). For

optical correlation, randomised groups of boreholes which are close to each other (e.g.,

Fig. 3. 21b: Kag_59, Kag_17, Kag_09 and Kag_11) were selected and, for each individual

borehole, a histogram of original data was constructed. The areas between selected

boreholes in the cross-sections indicate estimated Cu values. There is a general belief

that when the Cu values of the original data which are shown in the form of histogram for

each borehole (Fig. 3. 21) are high, the estimated Cu values within those two boreholes

should be high (RockWorksTM 15, 2010; Emery, 2012). For example, the Cu values of the

original data from deeper parts of boreholes Kag_17 and Kag_09 reveal high values of

element concentration therefor, the estimated Cu values between those original data

should consider high value of the element concentrations, as depicted in Fig. 3. 21b.

However, this method is not that sufficient because most of the boreholes were obliquely

drilled.

There are so many interdependent subjective decisions in a geostatistical study that it is

good practice to validate the results obtained by the estimation method (IDWAM in this

thesis) prior to any production run. The generated block model is validated by re-

estimating known values under implementation conditions, including the variogram

133
model, estimation method and search strategy, as close as possible to those of the

forthcoming production run (Deutsch and Journel, 1998). These re-estimation techniques

are discussed in most practical statistics and geostatistics books (Tukey, 1977; Efron,

1982; Davis, 1987; Isaaks and Srivastava, 1989; Goovaerts, 1997). The term jackknife

applies to resampling without replacement, i.e., when alternative sets of data values are

re-estimated from other non-overlapping data sets (Efron, 1982). The jackknife analysis

in the Kahang deposit indicates that the correlation between original data and Cu

estimated is 70%, as depicted in Fig. 3. 22. In this figure, the diagonal of the square plot

(black line) and the linear regression (red line) was derived and calculated using MATLAB

program.

(a)

134
 (b)

Fig. 3. 21. Optical correlation (visual verification) between Cu values of borehole and block
model

135
 Fig. 3. 22. Correlation chart between original and estimated data using jackknife resampling

3.7 Delineation of Mineralised Zones Using C-V Modelling

3.7.1 C-V Fractal Model
According to a study by Everets and Mandelbrot (1992), the relationship between the

measured parameters (e.g., ore grade, density and RQD) and numbers of 2D or 3D

dimensional cells can be easily determined. Partition function is addressed as follows:

N ( )

 q ( )  
i 1
q
i

Equation 3-6

136
Where q,  and  are the partition function, 2D or 3D blocks dimensions and a parameter

which is in investigation (e.g., ore concentration), respectively. If the investigated

parameter has the multifractal nature/behaviour in itself, the following relationship is

established:

 q ( )    ( q )

Equation 3-7

In this case, the partition function is equal to the exponential relationship between the

value of  which is q at any given moment.

Cheng et al. (1994) proposed the fractal Concentration–Area (C–A) model for separating

geochemical anomalies from background values in order to characterise the distribution

of major, minor and trace element concentrations in relation to the Mitchell-Sulphurets

porphyry system in British Columbia (Canada). This model has the general form:

A(ρ≤υ) ∞ ρ−a1; A(ρ≥υ) ∞ ρ–a2

Equation 3-8

where A(ρ≤υ) and A(ρ≥υ) denote areas (A) with concentration values ρ that are,

respectively, smaller and greater than contour value ρ defining that areas υ represents

the threshold), which define those areas and a1 and a2 are characteristic exponents for

both criteria. In log–log plots of concentration contours versus areas, certain

137
concentration contours representing breakpoints in the plots are considered threshold

values separating geochemical populations in the data. Zou et al. (2009) applied the

fractal C–A model to characterise the vertical distribution of element concentrations in the

Qulong copper deposit, Tibet, western China. Fractal models have been used to identify

the vertical distribution properties of Cu concentration values in mineralised and non-

mineralised zones. Cheng (2007) described hydrothermal processes (for undiscovered

mineral deposits in Gejiu, Yunnan Province, China) in the Earth crust associated with ore

deposits, such as porphyry ore deposits, which are characterized by high metal

concentrations having fractal or multifractal properties.

Cheng et al. (1994) and Zou et al. (2009) have suggested that the fractal C–A model is

applicable in volume or can be extended to volumetric extensions because element

distributions in horizontal or vertical directions are in accordance with fractal models.

Concentration-Area is for recognition of anomalies clearly in the areas, namely a two-

dimensional environment. As a result, the equations of C-A fractal model can be

established in the forms of:

Ah(  )    D

Equation 3-9

Av (  )    D

Equation 3-10

138
Ah and Av indicate two areas containing grades within horizontal (Plans) and vertical (cross

sections) directions, respectively. This reveals that the elemental distribution can also

exist in 3D dimension. Different forms of the C–A model expressed in Equations 3-9 and

3-10 can be rewritten as:

Ah(ρ≤υ) ∞ ρ−a1; Ah(ρ≥υ) ∞ ρ–a2

Equation 3-11

Av(ρ≤υ) ∞ρ−a1; Av(ρ≥υ) ∞ ρ–a2

Equation 3-12

where Ah(ρ≤υ), Ah(ρ≥υ), Av(ρ≤υ) and Av(ρ≥υ) denote two areas with concentration values

smaller and values greater than the contour value ρ defining that area respectively for all

variables are the same as those in Equation 3-6, but the subscripts h and v denote areas

described in horizontal and vertical section directions, respectively.

In Equation 3-7, (q) is an auxiliary function. With respect to the Equation 3-7, the

multifractal range of the  and the power value of α (q) can be calculated using the

following relationships:

𝜕𝜏
𝛼(𝑞 ) = ; 𝑓(𝛼 ) = 𝛼(𝑞 ) − 𝜏(𝑞)
𝜕𝑞

Equation 3-13

139
The multifractal range of 𝑓(𝛼 ) is a continuous function between two values αMin and αMax.

If (q) is a linear function with fixed values of á = áMax or á = áMin, 𝑓(𝛼 ) will be constant

which means that multifractal property changes to mono-fractal. In both cases, the

relationship between concentration and volume has to be discussed in various forms. In

the format of mono-fractal in which 𝑓(𝛼 ) and α are fixed, for better understanding of the

relationship between concentration and its occupying volume (the volume containing

specific grade) in the way that a database considering various grade is generated in

accordance with a determined voxel size (ε), the following relationships are used where

𝑉(𝜀) represents containing volume of 𝜌(𝜀).

𝑉(𝜀) ∝ 𝜀 −𝑓(𝛼)+2
Equation 3-14

𝜌(𝜀) ∝ 𝜀 𝛼−2
Equation 3-15

With the removal of ε from the two Equations 3-14 and 3-15, the relationship of

concentration and volume in mono-fractal (bi-fractal) behaviour is defined as follows:

𝑉 (𝜌) ∝ 𝜌 [2−𝑓(𝛼)]/(𝛼−2)]
Equation 3-16

In the above equation, values of 𝑓(𝛼 ) and α vary between zero and two. If a multifractal

model followed by a continuous range for 𝑓(𝛼 ) values exist then, two scenarios will occur.

140
In the first case, the α value is allocated to áMin (lowest value) and in the second case α

maximum value is equal to áMax. As a result, to obtain the relationship between

concentration and volume an integral equation is used as follows:

𝛼
+
𝑉(𝛼 < 𝛼𝑚𝑖𝑛 ) = 𝑉(𝜌) = ∫𝛼 𝑚𝑖𝑛 𝐶𝜀 −𝑓(𝛼)+2 𝑑𝛼
𝑚𝑖𝑛

Equation 3-17

Where C, ρ and 𝑉(𝜌) denote constant value, ore grade and occupied volume. Finally, the

relationship between concentration and volume in the multifractal mode can be addressed

as:

𝑉 (𝜌) ≈ 𝑉 (𝑇 ) − 𝐶𝜀 −𝑓(𝛼)+2 𝑑𝛼
Equation 3-18

In a 3D block model, the above equation can be expressed as:

V(  )    D
Equation 3-19

Where 𝑉 (𝜌) is a volume which contains ore grades equal and higher than ρ and D is a

fractal dimension.

141
Consequently, the relationships defined in Equations 3-11 and 3-12 suggest that element

distributions in a volume also follow a fractal model. The proposed fractal C–V model can

be expressed, therefore, in the following general form:

V)ρ≤υ(∞ ρ−a1; V)ρ≥υ( ∞ ρ–a2

Equation 3-20

Where, V(ρ≤υ) and V(ρ≥υ) denote volumes (V) with concentration values (ρ) that are,

respectively, smaller and values greater than contour values (υ),which defines those

volumes and a1 and a2 are characteristic exponents. Based on this kind of

characterization, it is the assumed hypothesis that different zones in porphyry Cu deposits

have fractal properties, which can be described by power–law relationships between ore

element concentrations and volumetric extensions (Afzal et al., 2011 and 2012; Wang et

al., 2013; Coghill et al., 2014; Sun and Liu, 2014; Awadelseid et al., 2015). In log–log plots

of concentration contours versus volumes, certain concentration contours, representing

breakpoints in the plots, are considered threshold values separating geochemical

populations within the data. To calculate V(ρ≤υ) and V(ρ≥υ) enclosed by a concentration

contour in a 3D model, in this study, the original borehole data of ore element

concentrations were interpolated by using the geostatistical and inverse distance

weighted (IDW) method. The interpolated 3D block model was used for the purpose of

this study. Volumes V(ρ≤υ) and V(ρ≥υ) are equal to the unit volume of a voxel (or volume

cell) multiplied by the number of voxels with concentration values (ρ) that are,

respectively, smaller and greater than a certain concentration value (υ). Log–log plots of

142
the concentration contours versus the corresponding volumes [V(ρ≤υ) and V(ρ≥υ)] follow

a power–law relationship.

Breaks between straight-line segments in those log–log plots represent threshold values

separating populations of geochemical concentration values. In typical porphyry Cu

deposits, which are mostly high tonnage – low grade, zones of high Cu concentrations

comprise relatively few voxels in a 3D block model, whereas zones of low Cu

concentrations comprise numerous voxels. Therefore threshold values in this are

recognised by applying the proposed fractal C–V model that likely represents boundaries

between different ore zones and ‘barren’ wall rocks.

3.7.2 Application of C-V Fractal Modelling

The C-V fractal model for Cu has been created according to the Cu 3D block model.

Threshold values were identified from the log–log plot in Fig. 3. 23, which demonstrates

a power–law relationship between Cu concentrations and volumes occupied. It reveals

that there are five populations according to the log-log plot corresponding to 0.075 wt.%,

0.42 wt.%, 1.86 wt.% and 3.2 wt.% Cu in the deposit (Fig. 3. 23 and Table. 3. 8).

Depicted arrows in the log–log plot indicate threshold values (e.g., breakpoints: Fig. 3.

23). These separate various straight line segments in the log–log plots. There is a sudden

change in the rate of decrease of the volume enclosed by high values of Cu. The first

threshold of 0.075 wt.% represents the beginning of the Cu mineralisation in this scenario.

As a result of this, the range of Cu concentrations less than 0.075 wt.% is deemed as

143
‘barren’ host rock including weakly mineralised zone within the deposit which may be

assumed as waste from an exploitation point of view (e.g., propylitic alteration zone in

Sungun porphyry copper deposit, NW Iran: Lowell and Guilbert 1970: Sim et al., 1999;

Berger et al., 2008; Asghari et al., 2009; Wang et al. 2013; Soltani et al., 2014; Zuo and

Wang, 2015). In addition, the first threshold value obtained by the C-V fractal modelling

(Cu = 0.075 wt.%) can be defined as a Cut-off Grade (COG) of the deposit for the

preliminary optimisation study but, the selection of an optimised COG is a dynamic

process (See table 5. 3 of the chapter five for calculation of an economic COG),

dependant on ore grade distribution, deposit geometrical shape and especially

economical parameters (e.g., Krautkraemer, 1988; Ataei and Osanloo, 2003; Gu et al.,

2010). However, what is classed as ‘barren’ today may be economic in the future

particularly when there is an increase in the metal world commodity price (Caccetta and

Giannini, 1988; Dagdelen and Mohammed, 1997; Cairns and Shinkuma, 2003; He et al.,

2009). The second threshold value of Cu is 0.42 wt.% where the main Cu mineralisation

starts. The range of Cu concentrations higher than 1.86 wt.% illustrates an enriched zone

for Cu. For these Cu concentrations the slope of the straight line fit is near to 90°.

Based on the 3D model of Mo distribution, volumes corresponding to different Mo grades

were used to generate a C-V fractal model. Threshold values of Mo were identified in the

C-V log-log plot which revealed five geochemical populations and four threshold values

equal to 13, 100, 316 and 645 ppm Mo in the deposit (Fig. 3. 23 and Table. 3. 8). Enriched

mineralised zones are deemed to have higher than 645 ppm because with these Mo

concentrations the slope of the straight line fit is close to 90°.

144
The main Mo mineralisation starts from the second threshold which is 100 ppm in this

kind of scenario. It is important to bear in mind that the Mo concentrations which are

greater than 13 ppm represent the start of Mo mineralisation. Cu and Mo log-log plots

have a multifractal nature for the elemental mineralisation in the deposit.

Table. 3. 8. Cu and Mo thresholds defined by the C-V model in the Kahang deposit

Geochemical Cu (wt.%) threshold Mo (ppm) Range Cu

Range Mo (ppm)
population value threshold value (wt.%)
First (Barren host
- - <0.075 <13
rock)
Second (Main
mineralisation 0.075 13 0.075-0.42 13-100
starting)
Third 0.42 100 0.42-1.86 100-316
1.86 (Enriched zone
Fourth 316 1.86-3.24 316-645
for Cu)
645(Enriched
Fifth 3.24 >3.24 >645
zone for Mo)

145
 Fig. 3. 23. C-V log-log plots: a) Cu, and b) Mo

3.8 Comparison and Correlation between Results of C-V Fractal and

Geological Modelling

To separate major mineralised zones including the supergene enrichment and hypogene

zones, a correlation between the geological model (as mentioned in section 3.7.2) with

Cu and Mo concentration distribution models has been constructed. In addition, results

146
from the C-V model were included in the combined model, where in consequence the

supergene enrichment zone exists in small sections close to the surface and its Cu

concentration value does not exceed that of 1.4 wt.% (Fig. 3. 24).

147
Fig. 3. 24. Geological zones (Cu distribution) including supergene enrichment (a) and hypogene
(c) with modified zonation models via C-V showing regions of supergene enrichment (b),
hypogene (d), main hypogene (e) and enriched hypogene (f)

148
Histograms of estimated Cu values for supergene enrichment and hypogene mineralised

zones show that there are two populations within the supergene zone, one of which

includes Cu values less than 0.2 wt.% and introduces weakly supergene mineralised zone

(Sillito, 1997; Pirajno, 2009; Fig. 3. 25a). Moreover, most parts of the hypogene zone

contain Cu values near to zero, as depicted in Fig. 3. 25b. It reveals that a correlation and

validation between results obtained by geological and C-V fractal modelling is essential

in order to achieve an accurate block model for a future optimisation study.

The supergene enrichment zone with Cu > 0.42 wt.% concentrations is located in a small

area within both the central and the eastern surface parts of the deposit, as depicted in

Fig. 3. 24. As can be seen, the supergene enrichment zone derived via the C-V model

has a volume smaller than its geological equivalent model.

The correlation between the geological hypogene zone and the C-V model indicate that

marginal parts of the geological model have Cu concentrations ≤ 0.075 wt.% and are

consequently considered as weakly hypogene mineralised zone. However, the main

hypogene zone with Cu ≥ 0.42 wt.% is located in the central, eastern and NW sections of

the deposit especially at depth, but in the NE part of the deposit it is close to outcropping.

The enriched hypogene zone with Cu ≥ 1.8 wt.% is situated in small sections of the

central, NW, NE and SE parts of the deposit, as illustrated in Fig. 3. 24.

The Mo distribution model is correlated with the supergene enrichment and the hypogene

zones, as shown in Fig. 3. 26. The maximum concentration of Mo in the supergene

149
enrichment zone is 104 ppm and high values of Mo are situated in the hypogene zone.

The main Mo mineralisation with Mo ≥ 100 ppm in the hypogene zone correlates with the

main hypogene zone (Cu ≥ 0.42 wt.%). The enriched Mo zone with Mo ≥ 645 ppm is

located in the central part of the deposit and associated with the enriched hypogene zone

(Cu ≥ 1.8 wt.%), as shown in Fig. 3. 26. These results indicate that the enriched

mineralised zone is located within the hypogene zone especially in the central, NW and

NE sections of the deposit.

In order to validate the results from the C-V model, a comparison between the

mineralogical model (for chalcocite and chalcopyrite distributions: Fig. 3. 27a and b) and

the main mineralised zones with Cu ≥ 0.42 wt.% (Table. 3. 8 and Fig. 3. 23a) was

conducted. To do this, the chalcocite and chalcopyrite mineralogical units were

distinguished using a mathematical filter facility within the RockWorksTM software which

is called “Boolean data type”. As a result, the studied mineralogical units in the 3D model

(See chapter 2, Fig. 2. 14c for the 3D dominant ore minerals) were allocated with binary

codes (zero or one). Consequently, zones with the code number of zero are removed and

the zones with the code number of one will remain in the 3D models (Fig. 3. 27a and b).

Subsequently, another mathematical facility within the software called “Multiple of Model

& Model” was used, which is a tool to manipulate the voxels in a solid model by the

corresponding voxels in another equally-dimensioned solid model file between 3D

mineralogical model and corresponding mineralised zones (e.g., RockWorksTM 15, 2010;

Fig. 3. 27c, d and e). After doing that, it has been shown that chalcocite is associated with

the supergene enrichment zone (Fig. 3. 27a and c) and chalcopyrite is also located within

150
the main hypogene zone with Cu ≥ 0.42 wt.% (Fig. 3. 27b and d). In addition, the

chalcopyrite from the mineralogical model containing Cu ≥ 0.42 wt.% has a proper

correlation with the main hypogene zone, as depicted in Fig. 3. 27e.

151
Fig. 3. 25. Histograms of estimated Cu values within the: a) supergene enrichment, and b)

hypogene zones

152
Fig. 3. 26. Mo distribution in supergene enrichment zone (a), hypogene zone based on Mo C-V
model (b), hypogene with Mo > 100 ppm (c), hypogene with Mo > 316 ppm (d) and Mo enriched
zone (e)

153
Fig. 3. 27. Correlation between chalcocite (a), chalcopyrite (b) and chalcopyrite ≥ 0.42 wt.% Cu
(e) zones with supergene enrichment zone (c) and main hypogene zone (d) based on C-V
model

154
3.8.1 Application of Logratio Matrix

Carranza (2011) has proposed a logratio matrix for further calculation of spatial

correlations between two binary models. Using the mineralisation model, an intersection

operation between a fractal mineralised zone model and different zones in the geological

ore model was performed so as to obtain the numbers of voxels corresponding to each

of the four classes of overlap zones as shown in Table. 3. 9 (the obtained values are the

overlap number of voxels between two binary geological and fractal models). Using the

obtained numbers of voxels, Type I error (T1E), Type II error (T2E), and overall accuracy

(denoted as OA) relate to the ability of the analysis to define ‘barren’ host rocks

(background) and mineralised zones delineated using the C-V fractal model with respect

to the geological models. Type I error (denoted as T1E) relates to the ability of the analysis

to define ‘barren’ host rocks whereas Type II error (denoted as T2E) relates to the ability

of the analysis to define mineralised zones. The lower the error (i.e. the higher value for

OA) the better the ability of the analysis to define ‘barren’ host rocks and mineralised

zones. The values for OA of the C-V fractal and geological models (alteration and

hypogene models) were compared with one another as follows:

155
Table. 3. 9. Matrix for comparing performance of fractal modelling results with geological model.
A, B, C, and D represents the number of voxels overlapping between classes in the binary
geological model and the binary results from fractal models (Carranza, 2011)

Comparison between the hypogene zone obtained from the geological model and the

main Cu and Mo mineralised zones from the C-V fractal model demonstrates that the

hypogene zone has a better correlation with the main Cu mineralised zone (Cu > 0.42

wt.%) because the number of overlapping voxels (A) in the main Cu mineralised zone

obtained using the C-V model (20,839 voxels) is higher than in the main Mo mineralised

zone (16,990 voxels), as depicted in Table. 3. 10. The overall accuracy of the main Cu

and Mo mineralised zone derived via the C-V fractal model with respect to the hypogene

zone of the geological model is equal to 0.154 and 0.146, respectively.

Alterations play a fundamental role in zone identification and also in presenting geological

models, as described by Lowell & Guilbert (1970). Correlation (from OA results) between

the main Cu mineralised zone obtained from C-V model and potassic alteration is higher

than the phyllic alteration because the OA for potassic and phyllic alterations have been

determined as 0.765 and 0.509 respectively (Table. 3. 11). As a result, the higher values

156
for overall accuracy in Tables 3. 10 and 3. 11 represent the higher overlap between

geological zones with mineralised zones identified by the C-V fractal model.

Table. 3. 10. Overall accuracy (OA) with respect to hypogene zone as delineated in the
geological model and Cu and Mo main mineralised zones obtained through C-V fractal model
(Values are the number of voxels)

Hypogene zone of Geological model

Inside zone Outside zone

C-V fractal model of Cu Inside zone A 20839 B 3348

main mineralised zone Outside zone C 411164 D 54576

OA 0.154

Hypogene zone of Geological model

Inside zone Outside zone

C-V fractal model of Mo Inside zone A 16990 B 1795

main mineralised zone Outside zone C 414954 D 54674

OA 0.146

157
 Table. 3. 11. Overall accuracy (OA) with respect to potassic and phyllic alteration zones and
Cu main mineralised zones obtained through C-V fractal model (Values are the number of
voxels)

Potassic alteration zone of Geological model

Inside zone Outside zone

C-V fractal model of Cu Inside zone A 2874 B 21313

main mineralised zone Outside zone C 93484 D 372256

OA 0.765

Phyllic alteration zone of Geological model

Inside zone Outside zone

C-V fractal model of Cu Inside zone A 10345 B 13842

main mineralised zone Outside zone C 226246 D 239494

OA 0.509

Validation between the main Mo mineralised zone (Mo > 100 ppm) based on the C-V

fractal model and alteration zones from the geological model indicates that there is a

difference between the two alteration zones. Overall accuracy for the potassic and phyllic

zones has been determined as 0.770 and 0.524 respectively (Table. 3. 12). According to

these results, the main elemental mineralised zones have better correlation with the

potassic alteration zone.

158
 Table. 3. 12. Overall accuracy (OA) with respect to potassic and phyllic alteration zones and
Mo main mineralised zones obtained through C-V fractal model (Values are the number of
voxels)

Potassic alteration zone of Geological model

Inside zone Outside zone

C-V fractal model of Mo Inside zone A 1699 B 17086

main mineralised zone Outside zone C 95053 D 374575

OA 0.770

Phyllic alteration zone of Geological model

Inside zone Outside zone

C-V fractal model of Mo Inside zone A 11531 B 7254

main mineralised zone Outside zone C 224919 D 244709

OA 0.524

3.9 Results
Results from this study show that the C–V fractal model can be used to recognise different

mineralisation zones in porphyry Cu deposits. Different geochemical populations can be

interpreted via the C–V fractal model. The C–V fractal model uses relationships between

the ore element concentration and the enclosing volumes, for example the concentration

of Cu associated with different zones, and satisfies power–law relationships. The

proposed fractal model could be applied for delineating enrichment zones from the

‘barren’ host rock, or from the background value using the concentration values of the

zones in combination with characteristic features of their geometrical shapes. The

159
proposed model is applicable to ore elements in the Cu and Mo porphyry deposits for

which the spatial patterns of concentration values satisfy a multifractal model.

Results from this study reveal that the hypogene zone is a major mineralised zone within

the Kahang Cu-Mo porphyry deposit. Based on the C-V fractal model, the threshold value

of 0.075 wt.% Cu may be equal to the deposit COG however, further investigation in this

regard has to be conducted considering economic aspects of the optimisation study (See

chapter 5). The threshold values for the main Cu and Mo mineralisation are 0.42 wt.%

and 100 ppm, respectively. Enriched Cu-Mo mineralised zones with Cu ≥ 1.8 wt.% and

Mo ≥ 645 ppm are located in the central, NW and NE parts within the hypogene zone.

The supergene enrichment zone exists in small parts within the deposit, especially in the

central and eastern zones.

The supergene enrichment and hypogene zones delineated by the C–V model correlate

well with the alterations and mineralogical data shown in the 3D models. The C–V log–

log plots from the Kahang deposit show that there is a multifractal model for Cu and Mo.

From a comparison of the C–V and geological models the supergene enrichment zone

shows a spatial correlation with the chalcocite-rich zone within the Kahang deposit. The

main hypogene zone of Cu ≥ 0.42 wt.% has an association with the chalcopyrite

distribution model. According to the correlation between results driven by fractal modelling

and geological models using the logratio matrix, the Cu and Mo main mineralised zones

generated in the C-V fractal model have a strong correlation with the potassic alteration

zone with respect to the OA.

160
 CHAPTER FOUR. Rock Mass Characterisation
Utilising Fractal Modelling based on Density and
RQD Data

161
4.1 Introduction

Assessment of rock density and RQD are crucial aspects of mineral exploration, resource

modelling and mine planning with huge cost implications for the design and the mining of

each block (ore or waste tonnage). These are calculated using the dimension and density

of each block. The results are assessed to identify a final pit slope angle and consequent

pit stripping ratio (Lerch and Grossmann, 1965; Little, 2006; Grenon and Hadjigeorgiou,

2010). The tonnage of these blocks located in each extractive zone is determined on the

basis of the density zones (e.g., ore and waste; Hamdi and Mouza, 2005; Hustrulid and

Kuchta, 2006; Rafiee and Vinches, 2008; Yasrebi et al., 2011). Host rocks of porphyry

deposits consist of sub-volcanic bodies such as porphyritic quartz diorite, granite,

monzonite and quartz monzonite which are lithological units with high hardness (Hitzman

et al., 1992; Laznicka, 2005). Parameters such as density, hardness, porosity and fracture

frequency (number of fractures counted each meter) give additional indications of rock

mass characteristics within porphyry deposits (Dershowitz and Einstein, 1988; Meyer and

Einstein, 2002; Kalenchuk et al., 2006).

Numerical models in geosciences have been created and consequently utilised to better

interpret the variability of geological parameters such as lithology, ore-type, alteration and

mineralogy or for a better understanding of the different attributes such as density, rock

mass characterisation and RQD (Jinga and Hudson, 2002; Rafiee and Vinches, 2008;

Yasrebi et al., 2013b, 2014). However, the classical statistical methods for delineation of

populations from a background level would be for example, a histogram analysis, box

plot, summation of mean and standard deviation coefficients and median. These are not

162
considered overly accurate due to the fact that these statistical methods consider only the

frequency distribution of information while not paying attention to the spatial variability

(Boadu and Long, 1994; Ehlen, 2000). In other words, the classical statistical plots (i.e.,

histograms) are based on the data abundance distribution and cannot quantify the spatial

positions of parameters such as RQD and density (Baecher et al., 1977; Rouleau and

Gale, 1985; Villaescusa and Brown, 1990; Lu 1997; Rafiee and Vinches, 2008; Madani

Esfahani and Asghari, 2013). As a result, numerical modelling of rock characteristics is a

difficult task and requires 3D modelling for better interpretation of the problems found in

a mining operation such as rock discontinuities, planar failure, circular failure, wedge

failure and toppling failure (Zhang and Einstein, 2000; Lina and Kub, 2006; Yasrebi et al.,

2013b, 2014).

The earliest model regarding the quantitative description of in-situ block size distribution

(IBSD) was the Rock Quality Designation (RQD; See chapter 1, section 1.1 for the RQD

definition), which was proposed by Deere (1964). Priest and Hudson (1976) applied the

RQD method to scanline survey data with respect to an analytical relation between RQD

and the discontinuity frequency resulted from a scanline survey. A borehole or a scanline

are by their nature one dimensional. As a result, RQD values calculated using bore hole

data or a scanline survey are influenced by the orientation in which the measurements

are taken (horizontal or vertical) so the method does not consider calculation for the other

direction (Lu, 1999; Carvalho et al., 2007; Slob, 2010). In order to overcome the

dependence of RQD on orientation, Kazi and Sen (1985) proposed the use of the

Volumetric Rock Quality Designation (V. RQD) which is a three-dimensional parameter.

163
This encloses the proportion of the volume of intact matrix rock blocks, equal to or higher

than 0.001 m3 in size, which can be associated with the average volume of a matrix block

and the number of matrix blocks per m3. To do this, the V.RQD is calculated by summation

of the volumes of intact blocks divided by the total rock mass volume which is expressed

as a percentage. However, the proposed model is limited to the estimation of the average

block volume rather than the IBSD (Deere and Deere, 1989; Şen and Eissa, 1992; Lu,

1999). Palmstrom (1985) proposed different empirical equations to link Volumetric

Discontinuity Count for RQD (known as Jv RQD) data and linear fracture frequency. He

suggested that there is a correlation between the in-situ block size and Jv is represented

in a figure incorporating various measurements of the block size or degree of jointing

(e.g., density of joints, RQD, block volume and joint spacing). However, this model can

only estimate a rough upper and lower range of block sizes and therefore has restricted

practical applications (Şen and Eissa, 1991; Milne, 2007). Şen and Eissa (1992) derived

values for Jv for RQD and block volumes of different shapes such as bars, plates and or

prisms quantity. The proposed model provides a simple tool for rock engineers without

the need for recourse to theoretical calculations. However, the block volume in this model

is given in terms of average block size so it cannot describe the block size distribution

(Lu, 1999; Palmstrom, 2005).

Techniques for selecting a theoretical function to describe rock mass characteristics

based on RQD data are unsatisfactory, as mentioned above. As a result, an approach to

introduce an analytical model to help achieve a proper fitted curve for interpreting the

distributions of measured RQD combining with density data is needed (e.g., Tables. 4. 5

164
and 4. 6; Figs. 4. 8; 4. 10 to 4. 12). An accurate description of density and RQD distribution

within a deposit plays a significant role in any optimisation study (See chapter 5, section

5.5 and 5.7). In recent years, models based on fractal geometry as a nonlinear

mathematical method, proposed by Mandelbrot (1983), have been widely used in different

branches of earth sciences since various geological processes and even mining-based

issues like rock mass characterisation can be categorised by changes in fractal

dimensions resulting from analysis of the relevant data and desirable attributes (density

and RQD in this scenario). Therefore, fractal analysis has the abilities to identify the

differences within ore deposits especially in hydrothermal occurrences such as porphyry

Cu deposits (Agterberg et al., 1993; Cheng et al., 1994; Li et al., 2003; Cheng, Q., 2007;

Harati et al., 2013). In other words, differences of fractal dimensions in density and RQD

data can certify applicable criteria to identify rock mass characteristics.

Consequently, the aim of this chapter is to use the Density–Volume (D–V) and RQD-

Volume (RQD-V) fractal models, which demonstrate that there is a power–law

relationship between desirable attributes (e.g., density and RQD values within the

deposit) and their cumulative volumes occupying spatial positions, to delineate density

and RQD populations associated with the major rock types (e.g., porphyritic quartz diorite:

PQD, andesite: ANS and dacite: DAC: Figs. 4. 8; 4. 11) which occupy 93% of whole

deposit volume (Yasrebi et al., 2013b). The results derived via the RQD-V fractal model

are validated using the Deere and Miller rock classification (1966; Table. 4. 1) to test

whether the RQD-V fractal model defines the best result with respect to the volume of

voxels (blocks) located within the mentioned lithological units in order to predict an

165
optimised experimental final pit slope. Moreover, for validation purposes and better

understanding of rock characteristics in the studied deposit, the Log-ratio matrix proposed

by Carranza (2011) has been employed.

Table. 4. 1. Classification of Rock Quality Designation, Deere and Miller rock classification
(1966)

4.2 Statistical Characteristics

RQD and density histograms provide a means for quickly evaluating the range of density

and RQD values for a selected data set (See chapter 3, section 3.3, figures for the original

data sets used for RQD and density values) without creating a 3D solid model in order to

illustrate the highest, lowest, sum, or average data values. These histograms are used to

read a single column of data (RQD or density) from a data set to determine the frequency

or percentage of the total number of measurements for that variable/attribute that falls

within each user-defined grouping. Location of boreholes which have been analysed for

density and RQD values is shown in Fig. 4. 1. In this deposit, 33 and 14979 density and

RQD samples have been measured from 11 and 42 boreholes, respectively, out of a total

of 48 boreholes carried out in the deposit (Fig. 4. 1; See chapter 3, Tables. 3. 2 and 3. 3;

Fig. 3. 4c and d). The density and RQD distribution functions are shown in Fig. 4. 2 and

166
4. 3, and are not normally distributed, with averages of 2.68 t/m3 and 48%, for density and

RQD, respectively.

Fig. 4. 1. Location of boreholes sampled for density and RQD

167
Fig. 4. 2. Density histogram based on raw data for the Kahang porphyry deposit

Fig. 4. 3. RQD histogram based on raw data for the Kahang porphyry deposit

168
The multi-modal distribution for density (Fig. 4. 2) indicates that the deposit has high

values in terms of density especially in the major rock type of PQD occupying most parts

of the deposit. As a result, most density values are higher than the mean point which

increases the tonnage of each individual voxel and correspondingly produces a rise in the

total rock tonnage. There is the bimodal distribution for RQD (Fig. 4. 3). As a result, there

are two main populations with values of <25% and >90% for RQD which illustrates that

there are two classifications of rock qualities (poor and excellent) with respect to the

Deere and Miller RQD classification (1966). In addition, with respect to the RQD

histogram, the greatest frequency of the RQD data corresponds to the excellent rocks

within the deposit which will result in the highest stability of the final pit slope. The tonnage

of extractive blocks in the deposit is calculated on the basis of rock density (ore and

waste) which is used for pit optimisation study.

4.3 Methodology

Initially, a database was generated based on drill core data consisting of lithological units,

density and RQD values. Secondly, the database was entered into the RockWorks TM 15

software package to build 3D density and RQD block models (Fig. 4. 5 and 4. 6) utilising

IDWAM due to the lack of adequate density and RQD data and having a non-uniform

drilling grid. For RQD, an experimental variogram in horizontal (with lag spacing of 15 m)

and vertical (with lag spacing of 8 m) orientations was produced via MATLAB software

with respect to log transformation of RQD values (raw data), as shown in Fig. 4. 4. From

this, the horizontal (Azimuth: 0 and Dip: 0) and vertical (Azimuth: 0 and Dip: -90) ranges

for RQD are 120 m and 270 m, respectively. However, no proper variogram can be fitted

169
for density because of so few data (just 33 samples: See chapter 3, Table. 3. 2 and Fig.

3. 4c). The next step was to propose the Density–Volume (D–V) and RQD-Volume fractal

models to generate the separation of different populations in terms of density and RQD.

Subsequently, a mathematical facility within software called “Multiple of Model & Model”

was used to manipulate the voxels in a solid model using the corresponding voxels in

another equal-dimension solid model. From this, a correlation between the density and

RQD 3D block models interpreted via D–V, RQD-V and the porphyric quartz diorite

lithological unit was conducted (e.g., RockWorksTM 15, 2010; Fig. 4. 12). The theoretical

variogram for RQD is as follow:

𝛾𝑅𝑄𝐷 (ℎ) = 135.71 + 190 𝑠𝑝ℎ (20, 20, 20) + 499.88 𝑠𝑝ℎ (120, 120, 270)

Equation 4-1

Fig. 4. 4. Experimental and theoretical variogram for RQD

170
 Fig. 4. 5. Density block model determined using estimated data

Fig. 4. 6. RQD block model in Kahang porphyry deposit determined using estimated data

171
4.3.1 D-V Fractal Model

Fractal and multifractal modelling was generated with respect to a partition function and

a power-law relationship between parameters such as density and RQD (Evertz and

Mandelbrot, 1992; Li et al., 2003; Carranza, 2008). The D-V fractal model was

constructed based on a Concentration-Volume (C-A) fractal model, shown in Equation 4-

2:

V(ρ≤υ)  ρ -a1 ; V(ρ≥υ)  ρ –a2

Equation 4-2

Where V(ρ≤υ) and V(ρ≥υ) denote two volumes with density values less than, or equal to,

and greater than, or equal to, the contour value ρ; υ which represents the threshold value

of a population (or volume); and a1 and a2 which are characteristic exponents. Threshold

values in this model indicate boundaries between different density populations within ore

deposits. To calculate V(ρ≤υ) and V(ρ≥υ), which are the volumes enclosed by a contour

level ρ in a 3D model, the drill core data of density values was interpolated by using a

geo-statistical estimation. The density 3D model was evaluated by IDW, which can

improve the interpolation of voxel values that lie between data point clusters and can be

useful for modelling drill hole based data in different types of deposits (Lima et al., 2013).

D–V fractal model reveals that there are two populations according to the log–log plot,

one above and one below 2.7t/m3 within the deposit (Fig. 4. 7), indicating that the rocks

with correspondingly high density commence from this break point. Most parts of the

172
deposit have density values higher than 2.7t/m3, especially in the central part of the

deposit (Fig. 4. 5).

Fig. 4. 7. D–V log–log plot in the Kahang deposit

Results of the D–V model are correlated to the major lithological units of the deposit

consisting of porphyritic quartz diorite (PQD), dacite and andesite. High density rocks

(≥2.7 t/m3) defined by means of the D–V modelling are clearly correlated with porphyritic

quartz diorite defined by the 3D modelling of lithological drill core data (compare Fig. 4. 5

with Fig. 4. 8a). However, the high density rocks are also associated with andesite in the

marginal parts of the deposits, as illustrated in Fig. 4. 8b. The high density is not present

in the majority of dacitic rocks, as depicted in Fig. 4. 8c.

173
174
 Fig. 4. 8. Voxels with density ≥2.7 t/m3 within lithological units: a) porphyritic quartz diorite, b)

andesite, and c) dacite

Carranza (2011) has provided a method for determining the overlap correlations between

two binary models (See Chapter 3, section 3.8.1). An intersection operation between the

results from the fractal model and major lithological units was performed to obtain the

number of voxels corresponding to each of the four classes of overlap zones. Using the

number of voxels, Type I error (T1E), Type II error (T2E) and overall accuracy (OA) of the

fractal model were calculated with respect to the lithological model. Correlation between

the high density rocks obtained via the D–V fractal model and the major lithological units

show that the porphyritic quartz diorites have higher OA compared with andesitic and

dacitic rocks (OAPQD=0.65, as shown in Table. 4. 2). Moreover, overall accuracies of the

175
high density zones with andesite and dacite are 0.35 and 0.32, respectively, as depicted

in Tables 4. 3 and 4. 4.

Table. 4. 2. Overall accuracy (OA), Type I and Type II errors (T1E and T2E, respectively)
with respect to PQD rocks resulted from geological model and high density rocks obtained
through D–V fractal modelling of density data (the obtained values are the overlap number
of voxels between two binary geological and fractal models)

Table. 4. 3. Overall accuracy (OA), Type I and Type II errors (T1E and T2E, respectively) with
respect to andesite rocks resulted from geological model and high density rocks obtained
through D–V fractal modelling of density data (the obtained values are the overlap number of
voxels between two binary geological and fractal models)

176
 Table. 4. 4. Overall accuracy (OA), Type I and Type II errors (T1E and T2E, respectively) with
respect to dacite rocks resulted from geological model and high density rocks obtained through
D–V fractal modelling of density data (the obtained values are the overlap number of voxels
between two binary geological and fractal models)

4.3.2 RQD-Volume (RQD-V) Fractal Model

The RQD–V fractal model which is developed based on Concentration-Volume (C-V)

fractal (Yasrebi et al., 2013b) for separation of rock populations based on RQD as an

important parameter for the rock mass characterisation, can be shown in Equation 4-3:

V(ρ ≤ υ) ρ–a1; V(ρ ≥ υ) ρ –a2

Equation 4-3

Where V(ρ ≤ υ) and V(ρ ≥ υ) denote two volumes with RQD values less than or equal to

and greater than or equal to the contour value ρ; υ which represents the threshold value

of a volume; a1 and a2 which are characteristic exponents. Threshold values in this model

represent boundaries between different rock mass populations of mineral deposits. To

calculate V(ρ ≤ υ) and V(ρ ≥ υ), which are the volumes enclosed by a contour level ρ in a

3D model, the borehole data of RQD values were interpolated by using the IDW

estimation. According to the RQD 3D block model, volumes corresponding to various

177
RQD values were calculated to derive a RQD-V fractal model. Threshold values of RQD

were recognised in the RQD-V log–log plot (Fig. 4. 9) which reveals a power-law

relationship between RQD values and volumes occupied. Depicted arrows in the log–log

plot illustrate threshold values at three breakpoints corresponding to 3.55%, 25.12% and

89.12% for RQD. Based on the log–log plot, the excellent RQD populations are

considered to have > 89.12%. The range of RQD values between 89.12% and 25.12%

indicate a combination of good, fair and poor rock mass quality of which definition is in

accordance with the Deere and Miller rock classification (Table. 4. 1 and 4. 5). However,

very poor rock characterisation is for RQD < 25.12% containing of threshold value equal

to 3.55% so therefore, there are two very poor RQD populations in this deposit

considering RQD-V fractal modelling.

Fig. 4. 9. RQD-V log–log plot in the Kahang deposit

178
 Table. 4. 5. RQD populations (zones) based on three thresholds defined from RQD-V fractal
model

Deere and Miller RQD RQD Range Obtained by RQD-N The Amount of Voxels in Each RQD
Range within the Whole Model
Classification Log-log Plot

Very Poor <3.55 111,697

Very poor 3.55 – 25.12 191,416

Poor, Fair & Good 25.12 – 89.12 180,524

Excellent >89.12 30,254

Based on the RQD-V fractal model, the majority of the deposit consists of very poor zones

which include 303,113 voxels (Fig. 4. 10 and Table. 4. 5). However, poor, fair and good

zones are present along NE-NW trend. Excellent zones in terms of RQD occur in the

central and NW parts of the deposit. As a result, for an RQD > 89.12%, the slope of the

straight line fit is near to 90° based on the RQD-V log–log plot (Fig. 4. 9).

179
 Fig. 4. 10. RQD populations within the Kahang deposit based on thresholds defined from the
RQD-V fractal model: a) very poor zones, b) very poor zones, c) poor, fair and good zones, and
d) excellent zones

180
Results from the RQD-V model were correlated to the major lithological units (e.g.,

porphyritic quartz diorite, andesite and dacite: Fig. 4. 11a, c and e) which were

constructed using RockWorks™ v. 15 software and drill core data. Rocks with excellent

RQD, defined using the RQD-V model, show a good spatial correlation with porphyritic

quartz diorites, defined by the 3D modelling, in comparison with the andesite and dacite

lithological units (Table. 4. 6), particularly in the central and NW parts of the deposit (Fig.

4. 11b). Therefore, it can be concluded that the porphyritic quartz diorite unit hosts

excellent RQD values.

Table. 4. 6. Amount of the excellent RQD populations (voxels) defined from the RQD-V fractal
model (RQD ≥ 89.12t/m3) in each major lithological unit within the Kahang deposit

The Amount of Voxels in Each

The Amount of Voxels in Each Major
Lithological Units Lithological Unit within the Whole
Lithological Unit Associated with
RQD≥89.12t/m3 (Excellent RQD)
Model

Porphyritic Quartz Diorite 394,122 15,085

Andesite 142,802 3,742

Dacite 122,384 3,687

181
182
 Fig. 4. 11. a) 3D lithology model for porphyritic quartz diorite , b) porphyritic quartz diorite unit
based on the RQD-V model for RQD > 89.12%, c) 3D lithology model for andesite , d) andesite
unit associated with RQD > 89.12%, e) 3D lithology model for dacite , and f) dacite unit
associated with RQD > 89.12%

4.4 Comparison between D–V and RQD Models

The RQD parameter is an extremely useful indicator of rock mass quality, especially if

used alongside density interpretation (Harrison, 1999; Zhang et al., 2012). This can lead

to better understanding of rock quality in the study area because density variation within

a rock mass has a direct relationship with changes in geomechanical properties such as

porosity and permeability (Singh and Baliga, 1994). Demonstration and analysis of the

correlation between RQD and density can be used in optimisation studies for the

determination of the ultimate pit limit and mine planning. As can be seen in the RQD

histogram (Fig. 4. 3) for all of the lithological units, the RQD average is around 48%

183
representing a “moderate” quality for the studied rock mass. However, use of the average

RQD can be misleading for design purposes. The increased RQD in the porphyritic quartz

diorite is accompanied by a corresponding increase in the density of the same rock type.

As a result, there is a positive correlation between RQD > 70% with density > 2.7t/m3, as

depicted in Fig. 4. 12. Since RQD is a quality indicator for the rock mass and is associated

with factors such as strength, modulus of elasticity, coefficient of permeability and

different rock types, it can also be an indicator of potential slope stability. This suggests

that the porphyritic quartz diorite, which is the main host rock in this deposit, with a RQD

> 70% and a density > 2.7t/m3, would be associated with competent areas for potential

bench and slope stability which will ultimately influence the future pit slope design and

ultimate pit limit.

Fig. 4. 12. Correlation between RQD > 70% with density > 2.7t/m3 block model within the

porphyritic quartz diorite

184
4.5 Results

From the D–V log–log plot for the Kahang deposit, there is a mono-fractal model with a

break point in density at 2.7 t/m3. Correlation between the results of D–V and RQD models

reveals that the rock units with a higher density (>2.7t/m3) are associated with RQD values

> 70%. The final pit slope geometry and ultimate pit limit will depend on the economic

evaluation of the orebody. However, an awareness of the spatial variability of parameters

such as RQD, density can be used to assess geotechnical characteristics of the rock

mass1. This can then be used to evaluate potential slope stability and be incorporated

into a geotechnical risk model for the final pit geometry. Regions of high RQD may be

targeted as offering greater potential for increased slope angles or locations for siting of

critical haul roads. Regions of lower RQD should, where possible, be avoided for final pit

limits as they will require lower slope angles. As a result, it seems rather obvious that

there are likely to be multiple populations, presumably related to geology, e.g. lithology.

Certainly from a slope stability point of view it would be expected that anyone examining

this data would consider at least multiple domains for slope stability assessments, hence

two final pit slopes would be selected in determining an ultimate pit limit. Further

geotechnical characterisation will, however, be necessary to establish any potential

influence of the 3D fracture network and presence of any major discontinuity- controlled

instability.

The threshold RQD value for excellent rocks is 89.12% based on the fractal model as

situated in the central and NW parts of the deposit. Models of good and fair rocks in the

1
See Appendix. E for Density and RQD plans in different levels

185
central, eastern and NW parts of the deposit contain 25.12%–89.12% RQD values

according to the RQD-V model. According to the correlation between results derived by

fractal modelling and the major lithological unit of PQD in the Kahang deposit, rocks with

excellent RQD defined by the means of RQD-V model have a sensible correlation with

porphyritic quartz diorites resulted from the 3D geological model.

186
CHAPTER FIVE. Mining Optimisation

187
5.1 Introduction
Pit limit optimisations form an integral part of open pit mine planning and combined with

the other mine planning tools such as pit design and cut-off grade (COG) determination

are used in open pit mine planning to define the final pit limit and open pit mining

sequences (Johnson, 1968; Hustrulid and Kuchta, 2006). Various definitions and

recognised techniques for the analysis of pit limit optimisation results have been

introduced, developed and consequently improved by the mining industry. Following this,

Armstrong (1990) said the Ultimate Pit Limit (UPL) is the maximum boundary of all

materials certifying the criteria of:

1. A block of the material will not be mined unless it can cover all costs for its

subsequent mining (ore and waste), processing (ore) and marketing (ore).

2. Any block meeting the first criterion will be included in the pit.

Whittle (1988) suggested that an ultimate pit meets the highest possible undiscounted

cash flow without considering of scheduling target including pushbacks and consequent

mine planning. Hustrulid (1995) proposed that the pit remaining at the end of mining (mine

life) is called the final or ultimate pit. However, the destination of the material with different

money values defined as cut-off grade must be identified meeting economic criteria. In

total, an ultimate pit is the pit producing the highest value of NPV compared with the other

potential pits. One of the best ways to recognise a final pit limit was suggested by Lerch

and Grossman whose 3D graph theory is a practical computer-supported alternative

(Dynamic programming) to the conventional manual approach for open pit design through

a block model. They introduced a block model of a mine by a weighted directed graph

188
where each vertex represents the blocks and each arc represents the blocks

interdependency from an excavation perspective. The direction of the arcs from a vertex

to the other vertex reveals the excavation priority of the second block to the first block

and so on by the weights (e.g. NPV of each block) which comes from the blocks’ economic

values (Caccetta and Giannini, 1988; Akbari et al., 2008; Yasrebi et al., 2011). They

proposed that the aim of an ultimate pit limit is to find the maximum weight of the above-

stated weighted directed graph. In other words, the most famous optimisation algorithm

is the Lerch and Grossman algorithm which considers the influences of operating costs,

product prices, slope geometry, etc. The Lerch and Grossman algorithm is utilised with

varying revenue factors to develop a value mining-based sequence strategy which

introduces pit shells from revenue factors. The actual design has to also address

functional considerations such as haul road access, cut-off grades and processing, etc.

The determination of the final pit limit is one of the most significant aspects which must

be frequently reviewed and correspondingly corrected in the early stages and throughout

a mine life on the basis of deposit information and changes in economic parameters due

to uncertainty of the relative metal’s world commodity price and related mining costs

(Dimitrakopoulos et al., 2002; Akbari et al., 2008; Yasrebi et al., 2011; Asad and

Dimitrakopoulos, 2013). As a result, ultimate pit recognition in each period of time is a

function of financial affairs, which is well defined by Break-Even Stripping Ratio (BESR).

The calculation of economic elements of a deposit therefore has to be performed

according to the final exploration information and economical regime of the country in

which the project is being carried out (Johnson, 1968). In other words, identification of a

189
final pit limit can be examined at almost every stage of a project, from exploration program

definitions to the preparation of feasibility studies, modifications of development options

for an open pit mine and pit development sequence.

The determination of a final pit limit in an open pit mine in various implementation forms

is one of the most fundamentally important aspects of mine design which can produce

feasible optimum pit development geometries considering the geology, grade, slope and

economic information. Nowadays, in optimisation of open pit mines, determination of the

ultimate pit limit is just one of the many steps which are used in optimisation studies and

following this, engineers continue to utilise software packages which can achieve yearly

optimised mine plans or even selective mining designs. The goal of this chapter is to

determine the ultimate pit limit of the Kahang deposit by employing 3D block models (for

Cu and Mo) via the C-V fractal model and rock mass characterisation through D-V and

RQD-V fractal models by which the achieved results can be comprehensively adjusted

for all kinds of open pit mines in a way that can be used by mine planners.

One way to maximise the use of block modelling functions in order to optimise the pit

design process is to fully integrate block modelling and slope stability analysis. This is

because it is believed that optimised slope stability results in a lower amount of waste

material removed which reduces mining cost and correspondingly raises the NPV of the

whole project (Lerch and Grossmann, 1965; Koenigsberg, 1982; Hustrulid and Kuchta,

2006; Yasrebi et al., 2011 and 2014; Marcotte and Caron, 2013). In addition, there is a

logical action where one identifies different rock types (ore or waste) in terms of the

190
grades of each block. This process can be further enhanced by defining at every block

location an identified COG.

Before performing any of the computerised optimisation processes, a range of basic

information was required for the study (Fig. 5. 1). Technical data and economic

information are crucial within the optimisation process. These factors greatly influence pit

design. A sensitivity analysis was carried out on the basis of technical and economic

parametric changes for the mining strategy (See chapter 6). The Lerch and Grossman

algorithm considers the influences of operating costs, product prices, slope geometry,

etc.

(a)

191
 (b)

Fig. 5. 1. a) Design procedure in an open pit mine with regard to ultimate pit limit determination
(Akbari et al., 2008) and b) steps in mining design and planning by circular and interdependent
analysis (Osanloo et al., 2008b)

5.2 Methodology
Initially, the dataset obtained from a block model, via the C-V fractal model, was exported

in the form of a table or Open Database Connectivity (ODBC) format which is compatible

with optimisation software packages. Before this, careful validation of all data analysed

from boreholes was conducted, which is an important preliminary action before

generating a block model. The RockWorksTM 15 enables us to generate a 3D geological

and deposit block model which includes ore grade, rock density and rock type. The result

of this was a deposit reserve estimation and the likely modification of existing geological

maps matching the general observation of the region in question, and new cross sections.

Following this, NPV Scheduler was used in order to establish the final pit limit in terms of

the maximum Net Present Value (NPV) and associated ‘’Pushbacks’’ to produce a best

case mining scenario (Hustrulid and Kuchta, 2006; Elkington and Durham, 2011; Yasrebi

192
et al., 2011; Armstrong and Galli, 2012). To do this, all the required data such as grade,

density and rock type (1 for ore and 0 for waste) and other similar data were entered as

numerical values into each of the deposit’s block models. Resulting 3D models were

exported mathematically as 3D matrices into the optimisation programs in the following

manner:

1- Coordinates for the centre of sub cells in each plan were entered into an EXCEL

spreadsheet;

2- The layout of grade-based coordinates was adjusted using ACAD;

3- Coordinates and grades for each block (in all plans) were input into an EXCEL

spreadsheet;

4- Contours were drawn for minerals and specified zones in each plan;

5- Created blocks and the grade database were harmonised;

6- Overlaps existing in the blocks were removed;

7- Grades were allocated for each block and the geological 3D models were

controlled with high accuracy.

5.3 Deposit Block Model via the C-V Fractal Model for Optimisation
Study

Data from the C-V fractal model for the deposit (Fig. 5. 2) was entered into the

optimisation algorithms. As discussed in chapters 1 and 3 (See chapter 1 section 1.3,

chapter 3 section 3.7.2, Table. 3. 8 and Fig. 3. 23a), the C-V fractal model has provided

a mathematic tool to delineate geochemical populations from ‘barren’ host rock, including

193
weakly mineralised zones, thus providing a cut-off grade. The mineralised zones with Cu

less than 0.075 wt.%, apparent from the C-V log-log plot (Table. 3. 8 and Fig. 3. 23a), are

assumed to be waste materials which decrease NPV by increasing stripping ratio

(Hustrulid and Kuchta, 2006; Grenon and Hadjigeorgiou, 2010). From this, a Cu deposit

block model was generated, at Cu ≥ 0.075, resulting in lower stripping, as shown in Fig.

5. 2. As can be seen, the weakly mineralised zones with less than 0.075 wt.% Cu occur

within the marginal parts of the deposit (e.g., propylitic and argillic alteration zones; Lowell

and Guilbert,1970: See also chapter 2, Fig. 2. 5).

Fig. 5. 2. a) Estimated Cu block model and b) estimated Cu block model excluding Cu ≤ 0.075

wt.%, generated using the C-V fractal model

5.4 Mine Topographical Features of Land Surface

Topographical features of the deposit land surface (Fig. 5. 3), as well as other related

data, are presented in a 3D block model entered in the optimisation software, prepared

194
using the most recent topographical maps. It is clear that blocks located between

topographical surfaces and deposit surfaces are considered as waste blocks and are

entered in the economic model as blocks of negative significance.

Fig. 5. 3. Land topographical surface of the deposit for optimisation study

5.5 Pit Geometrical Characteristics

The geometrical parameters used in the pit design will now be discussed. Some of these

parameters are approximate. If they are changed then new optimisation studies have to

be carried out. However, generally, the variance of the results obtained from changing

these parameters are so insignificant that many of these parameters could be used

reliably with current approximations. Final slope angles of the mine are considered to be

195
the most influential geometrical parameters in an optimisation study (Singh and Baliga,

1994; Wyllie and Mah, 2004; Hustrulid and Kuchta, 2006; Grenon and Hadjigeorgiou,

2010). Optimisation software algorithms have been designed in such a way that the

blocks of different levels are extracted given this gradient (final pit slope). In NPV

Scheduler, a slope region is a physical volume to which a particular group of overall slope

angles and corresponding azimuths are defined (Fig. 5. 4).

As a general rule, slope stability studies for establishing an accurate final pit slope should

be accomplished prior to optimisation studies. But to do such studies, some data is

needed which are often obtained after completion of drilling operations and geotechnical

studies (Little, 2006; Grenon and Hadjigeorgiou, 2010; Yasrebi et al., 2014). Therefore,

calculation of a final slope for a mine is the most essential matter which should be studied

carefully at the initial stages of design (See chapter 4). Unfortunately, due to the lack of

systematic drilling and sub-surface data in the Kahang deposit, the possibility of studying

slope stability is precluded. So, error percentage of any study performed using these

kinds of observations in this regard will be very high and accordingly the accuracy of the

determination of stable slopes could not be guaranteed. To determine the gradient,

Density-Volume and RQD-Volume fractal models have been created. As a result, the

mine’s general gradient of 35° and 45° have been applied for performing optimisation

calculations (Fig. 5. 4). However, the extracted benches are characterised according to

the general features of the mine under consideration and more importantly on the basis

of extraction capacity and the machines and equipment to be allocated during mining. In

consideration of the current imposed constraints on the application of heavy machines as

196
well as harmonisation of the plan and executed plans, the heights of the benches have

been constrained to 10 m.

Fig. 5. 4. Stable pit slopes in the Kahang deposit, input into NPV Scheduler (See also chapter 4)

The results obtained from the optimisation software naturally have a ‘blocky’ nature so it

is necessary to use the characteristics of the benches and the roads in detail for

optimisation within the supplied ultimate pit surface. Given that, the phases and then

pushbacks need to be designed at different working slope angles to the final overall slope

angles. It makes more sense to use the working angles for internal phase and subsequent

pushback development (NPV – Scheduler. 2001). Given the previous experiences of

neighbouring deposits (e.g., Sharif Abad Cu porphyry deposit located in Isfahan province,

central Iran; Dareh-Zereshk and Ali Abad Cu porphyry deposits in Yazd province and Dalli

Cu porphyry deposit located in Markazi province), these characteristics have been

determined as follows:

197
  Face angle of the extracted benches: 70°

 Width of interconnecting roads: 10m

 Gradient of interconnecting roads: 8% (Max)

The estimate of cut-off grade is not consistent with the corresponding values for other

deposits in the region because it is dependent on ore grade distribution, deposit

geometrical shape and especially economical parameters which are different from one

deposit to another (Lerch and Grossmann, 1965; Lane, 1988; Osanloo and Ataei, 2003;

Hustrulid and Kuchta, 2006; See Chapter 3, section 3.7.2 and Appendix. A).

5.6 Mine’s Annual Production

The mine’s annual production capacity is one of the factors determined using economic

parameters and project profitability studies. Similarly, determination of production

capacity in this research should be subject to various detailed and basic studies.

Fortunately, optimisation software has useful capabilities that make it possible for

designers to conduct such studies extensively. Since the annual production will be limited;

i.e. around 1,000,000 tonnes (as the nominal capacity of the plant); this amount has been

exactly included in the calculations.

5.7 Ore Density

The cost of mining each block has been calculated using the dimension and specific

density of that block. It is clear that the tonnage of blocks located in each zone is

determined on the basis of each zones’ ore density. The average density of three major

198
lithological units of porphyritic quartz diorite (PQD), andesite (ANS) and dacite (DAC)

within the Kahang deposit is shown in Table. 5. 1.

Table. 5. 1. Ore density average of the mineralisation zones within the deposit

Density Average
Lithological Unit
(t/m3)
Porphyritic Quartz Diorite 2.67
Andesite 2.69
Dacite 2.68

The density 3D model for the studied deposit, including all densities within the lithological

units (three of which are illustrated in Fig. 5. 5), were generated using IDWAM. The aim

was to generate the finalised block model (database) for the optimisation study (Yasrebi

et al., 2014).

199
(a)

(b)

200
 (c)
Fig. 5. 5. Density distribution for block models in the Kahang deposit for a) PQD, b) ANS, and c)
DAC lithological units (See abbreviation table for lithological units)

5.8 Kahang’s Exploitation Percentage

To determine the amount of exploitable Cu and Mo in each tonne of ore, and so calculate

the value of each block, some factors such as processing efficiency within the different

ore minerals must be calculated and applied in this study. By referring to the results of

mineral processing tests, obtained by the laboratory, and the data sources (See

Appendix. A) provided by the deposit owner, NICICO, the percentage of mining recovery,

mining dilution and recovery fraction for concentrated Cu and Mo have been specified as

follows:

 Mining recovery: 95%

 Mining dilution: 3%

 Recovery fraction for Cu and Mo: 80%

201
5.9 Economic Principles
5.9.1 Prices and Expenses
The results obtained from optimisation studies are significantly affected by the price of

the product (Fig. 5. 6) and operational costs of production. These parameters must

therefore be determined more precisely. However, due to the importance of economic

principles, and the large influence they have on the results, the above-mentioned

parameters are discussed in detail in the subsequent sections.

Determination of operational costs including ore and waste exploitation costs and milling

costs for the Kahang deposit is a demanding task. However, after consideration of studies

conducted previously and simulation of cost figures incorporated in the records of Iranian

(as well as western authorities; See chapter 1, section 1.5 for political context of mining

in Iran), these values were determined by the deputy of the Economic and Financial

Department of NICICO (See Appendix. A for data sources) and then included in the NPV

Scheduler software (Table. 5. 2).

202
 (a)

(b)
Fig. 5. 6. Metal commodity prices during the optimisation study: a) copper, and b) molybdenum
(London Metal Exchange, 2015a and b)

203
 Table. 5. 2. Prices and mining costs for the Kahang optimisation study

Additional

Milling
Mining Mining Milling
Unit Unit Cost Unit
Price Unit Cost CAF Cost
(refinery)

& Selling

Cu 7000 $/tonne 4 $/m3 1.3 4 $/tonne 200 $/tonne

Mo 0.023 $/g 4 $/m3 1.3 4 $/tonne 0.0002 $/g

Waste - - 4 $/m3 1 - - - -

*Rock type mining CAF (Cost Adjustment Factor) = Mining cost for rock type/ Reference

waste mining cost

Comments
**
The metal commodity price was considered on the date when the optimisation study was

conducted

5.9.2 Annual Discount Rate

During strategic optimisation calculations, valuation of the blocks will be carried out on

the basis of exploitation time. Indeed, this task was undertaken by consideration of an

annual discount rate and updating the value of the blocks intended to be exploited in the

coming years. By comparing the optimisation studies of different mining projects in Iran,

it can be concluded that this annual rate varies slightly from project to project as this was

defined by the Central Bank of Iran (CBI) from 10% to 18% for short-term and long-term

projects, respectively (Fig. 5. 7a; Appendix. A. However, discount rate can be higher or

less than the above-mentioned one for foreign investors in regard to their countries’

economic principles (e.g., in the United Kingdom, rates were cut to 0.5% by the Monetary

Policy Committee; Fig. 5. 7b). Iran has been a World Trade Organisation (WTO) observer

204
member since 2005 and policy of the Iranian government is to facilitate absorption and

development of foreign investment for the country’s mining projects as well as oil and gas

development projects by providing a secure investment climate and creating free trade

zones (See chapter 1, section 1.5). In addition, because of low labour and energy costs

in Iran, mining projects and related minerals production, which consume huge volume of

energy, are profitable for foreign fund managers and companies (e.g., British-based

mining corporations, multinational-based mining companies and banks).

(a)

205
 (b)

Fig. 5. 7. Annual discount rates for choice of parameters and model outputs for domestic and
foreign decision makers, a) Iran b) United Kingdom (Trading Economics, 2015)

Given that the Kahang deposit is considered as a short-term project, a discount rate of

14% has been conservatively applied in optimisation calculations because the project

belongs to an Iranian based-corporation (NICICO). The project owner, as well as other

individual and governmental mining sectors in Iran, are seeking foreign investors, as

revealed by the Minister of Industries, Mines and Commerce of Iran (See section 1.5).

However, an extension of exploratory boreholes for the Kahang deposit is required, which

is likely to increase the reserve due to better geological constraints. The discount rate of

the project will be higher than 14% if the project expansion goes ahead which

correspondingly will increase the mine life so that Kahang becomes a long-term project.

The supplied discount rate is intended to determine the discounted cash flow (DCF). The

sequence used for allocation of blocks with respect to time is determined so as to

demonstrate the highest total DCF.

206
5.9.3 Cut-off Grade
Specifying the cut-off grade strategy is a matter which should be discussed before

studying the results of optimisation. Cut-off grade is usually calculated taking

consideration of the cost of mining, cost of processing as well as the price of the

commodity (Cairns and Shinkuma, 2003; Osanloo and Ataei, 2003; He et al., 2009; Wang

et al., 2010b). However, prior to performing such studies, it is necessary that a cut-off

grade needs to be considered for separating the blocks containing ore and waste. In

optimisation software, a grade is calculated similarly which is called the marginal limit

grade. The COG for the optimisation study is calculated in order to identify the best course

of action, either to mine or to leave, to mill or to dump. A COG of 0.07 wt.% Cu was

calculated based on Equation 5-1 as follows:

NPV = (I1+I2)-(MC+PC1+PC2) Equation 5-1

I1 = (G1×D×V×PCu×MR×PR)/100 Equation 5-2

I2 = G2×D×V×PMo×MR×PR Equation 5-3

MC = D×V×MCPU Equation 5-4

PC1 = (D×V×PCPU×G1×MR×PR)/100 Equation 5-5

PC2 = (D×V×MR×PR×PCPU×G2)/1000000 Equation 5-6

Where NPV, I1, I2, MC, MR, PC1 and PC2 are Net Present Value of each voxel, incomes of

Cu and Mo, mining cost for a voxel, mining recovery and processing costs for Cu and Mo,

respectively. Furthermore, PCu, PMo and PR are the metal commodity price for Cu and Mo

and fraction recovery. In addition, D, V, MCPU, PCPU, G1 and G2 denote voxel density,

voxel volume, mining cost per volume unit, processing cost per tonnage unit, Cu and Mo

207
concentration values, respectively. Economic cut-of grade was calculated using Equation

5-1, illustrated in Table. 5. 3. This indicates that the economic COG for Cu as the main

target is equal to 0.07 wt.% which has almost equal to the first threshold obtained by the

C-V fractal modelling (See Chapter 3, section 3.7.2).

Table. 5. 3. COG specification for optimisation study

Cu Density Volume Processing Cost

Mo (ppm) Mining Cost ($/m3) Cu Price ($/Tonne)
(wt.%) (t/m3) (m3) ($/Tonne)

0.07 17.309 2.25 160 4 200 7000

Mo Income Processing
Mining Processing Income Voxel Mining Processing Voxel
Price Cu Cost Cu
Recovery Recovery Mo ($/g) Cost ($/m3) Cost Mo ($/g) NPV ($)
($/g) ($/Tonne) ($/Tonne)

0.023 0.97 0.8 1368.864 111.215 1440 39.110 0.967 0

5.10 Determination of the Kahang Deposit Ultimate Pit Limit

5.10.1 Ultimate Pit Limit
Specifying the ultimate pit limit is the first step in optimisation studies. This limit is typically

specified by application of the Lerch & Grossman algorithm in the mine’s economic model.

Through this method, a pit with maximum cash flow is determined. If the magnitude of the

pit is greater than this limit, it means that profitability will be low. Tables 5. 4 and 5. 5

depict the imported model data (the database obtained by the means of the C-V fractal

model extended by topography) and economic model of the Kahang deposit (Yasrebi et

al., 2011; Marcotte and Caron, 2013).

208
Table. 5. 4. Initial imported data into the optimisation software based on Fig. 3. 13 and Table. 5.
2

Table. 5. 5. Economic specification of the Kahang deposit driven by NPV Scheduler

209
The individual economic parameters of all sequences, before a finalised optimisation

(1580 m<Elevation<2360 m), and some supporting 2D economic models in different

section levels are given in Appendix. F.

As a general rule, nested pits which have the greatest cash flow are considered as the

mine’s ultimate pit limit. Tables 5. 6 and 5. 7 represent the specifications of the final pit

and ultimate pit limit reserve report. As a result, the pit located at the elevation equal to

2225 m (Fig. 5. 8) is determined as the Kahang final pit limit in which the NPV value, ore

and waste amount, strip ratio and mine lifetime are $3,032,862, 3,648,294 tonnes,

13,970,954 tonnes, 3.8, 3.65 years, respectively (See Appendix. G).

Fig. 5. 8. Kahang pit limit 3D view without consideration of ramps and safety berms

210
 Table. 5. 6. Ultimate pit statistics determined by NPV Scheduler

Table. 5. 7. Pit limit reserve in the Kahang deposit

5.10.2 Internal Pit Shells (Phases)

The application of optimisation software will not be limited to determination of a pit’s

optimised limit, but after completion of this phase, a general scenario for exploitation of

this pit and so-called exploitation sequences (Nested pits) will be discussed and optimised

in the form of some smaller pits (Koenigsberg, 1982; Bond, 1995; Yamatomi et al., 1995;

211
Darwen, 2001; NPV – Scheduler, 2001; Goodwin et al., 2008; Marcotte and Caron, 2013).

Occasionally, a mine is exploited at one stage and the ultimate pit is not divided into

smaller pits. Indeed, this condition is considered as the worst mining scenario. To improve

the economic and technical outcome, the ultimate pit will be designed in the form of mining

sequences as nested pits. By increasing the number of mining phases, the best mining

scenario will be achieved due to increase the number of working benches increasing

production capacity. In a different definition, after generation of the ultimate pit, internal

phases (pit shells) will be produced (Table. 5. 8 and Fig. 5. 9), each of which may be

considered as an optimal pit corresponding to the 'worst-case’ economic scenario

compared with that derived using case-base parameters (economic principles).

The phase conducted for a supplied final pit is utilised as the basis for generation of a

nested pit, which is a block by block extraction sequence, and consequent optimal

extraction sequence (OES). However, each phase has been constructed to present the

highest undiscounted cash flow based on economic principles. Accordingly, specifying

the optimised number which ensures the project’s profitability and does not impose

restrictions executively is generally considered to be the most important fundamental of

designing open pit mines (Akbari et al., 2008; Yasrebi et al., 2011).

212
 Table. 5. 8. Pit optimisation phases of the Kahang deposit

Fig. 5. 9. Pit optimisation within the internal pit shells, stages of 80%, 90% and 100%, driven by
NPV Scheduler based on Table. 5. 8 (Cumulative profit, Incremental revenue, Incremental
total ore and Cumulative total ore)

213
5.10.3 Nested Pits
Nested pits are a collection of optimised pits which are calculated on the basis of the

Lerch & Grossman algorithm (Hustrulid and Kuchta, 2006). Indeed, constructing

optimised nested pits is considered as an ‘art’ applied by design engineers to perform

analyses such as specifying the ultimate optimised pit as well as determining the

extraction sequence of the blocks.

The total number of extraction sequences existing in the Kahang deposit reaches 100. In

other words, a maximum revenue factor of more than 100%, which is recommended in

NPV Scheduler software, can be determined, in which case the largest ultimate pit

produced is a pit ‘past the peak’, in terms of the base economic parameters (NPV –

Scheduler, 2001; See Table. H. 1 for both incremental and cumulative NPVs). Tonnage

of exploitable mineral, waste, waste ratio, cash flow and current value of the pits have

been calculated individually as depicted in Appendix. H. With this useful data, one is able

to perform the required analysis, detect a pit’s optimised limit and design an exploitation

schedule and timetable. In other words, the importance of nested pits is not always

evident in the long-term open pit planning procedure. Usually a mining sequence is

derived from a simple selection of pit shells based on optimum pit limits. The performance

of the obtained mining sequence to the production constraints is generally not questioned

prior to the detailed production stages of a project.

5.10.4 Identification of an Optimal Extraction Sequence (OES)

Identification of an optimal extraction sequence (OES) is significant for the selection of

alternative optimal pits in order to generate the best mining sequences (Dincer, 2001;

214
Godoy and Dimitrakopoulos, 2004; Hustrulid and Kuchta, 2006; Osanloo et al., 2008a;

Elkington and Durham, 2011). For all of the blocks inside the ultimate pit, an OES is also

constructed. This aims to achieve the highest DCF based on the given discount rate and

ore processing rate. The phase structure created for a supplied final pit is utilised as the

basis for constructing the OES (a block by block extraction sequence). It may be that

some blocks towards the end of the ultimate pit may not be added to the DCF, in which

case a smaller pit than the ultimate pit could produce the highest DCF. In other words,

the ultimate pit limit may have smaller magnitude than the recognised final pit when it

indicates that the addition of higher sequences does not increase the pit NPV (See

Appendix. H, rows 92 to 100 for cumulative NPV). By referring to Appendix. H, sequence

No. 92 (specified as the Kahang ultimate pit limit due to the fact that the NPV cumulative

trend becomes steady), the mining operation can be terminated at this point (Fig. 5. 10).

The exploitable reserve existing (ore) at this point (from sequence 1 to 92) amounts to

3,291,944 tonnes with the total NPV of $2,884,968 and a strip ratio of 3.919. Furthermore,

the cumulative profit value at the mentioned extraction sequence is equal to £7,853,825.

It is good to bear in mind that the data obtained from this pit should not be considered as

a basis for the design during optimisation studies because mining orientation has not

been yet recognised. However, after selecting and designing the extraction phases

(pushbacks), an optimised pit will be derived.

215
 Fig. 5. 10. Comparison between incremental and cumulative NPV values for the Kahang
deposit, driven by NPV Scheduler based on Table. H. 1 (the black arrow indicates sequence
No. 92 which specifies the Kahang ultimate pit limit)

5.11 Comparative Case Study

The boreholes drilled in the Kahang deposit are not evenly distributed (anisotropic grid

drilling), with a particularly large gap between the main cluster and the three drill holes

(KAG-43, KAG-38 and KAG-30) located in the NW of the study area, as depicted in the

Fig. 5. 11. The gap is due to the existence of a private garden that the National Iranian

Copper Industries Co (NICICO), as the project holder, is not allowed to enter and conduct

any drilling, even underneath. However, the aim of the comparative case study is to ignore

the three isolated boreholes and to compare the relative changes in terms of reserve

216
estimation, Concentration-Volume (C-V) fractal log-log plot and finally the consequential

variance to the NPV, with this mentioned scenario.

Fig. 5. 11. Disposition of boreholes in the Kahang deposit

As a result, the Kahang deposit was modelled with 263,410 voxels corresponding to
112,950,208 tonnes of sulphide ore (Fig. 5. 12) with an average grade of 0.166 wt.%
based on a Cu distribution function which is not normal (Fig. 5. 13).

217
 Fig. 5. 12. 3D Cu block model excluding the three isolated boreholes

Fig. 5. 13. Cu histogram from original data in the Kahang deposit excluding the isolated

boreholes

218
The C-V log-log plot for the new Cu block model indicates that there are four Cu

populations corresponding to 0.071 wt.%, 0.4 wt.% and 1.86 wt.% (Fig. 5. 14 and Table.

5. 9). Cu concentrations higher than 1.86 wt.% are from an enriched zone.

Fig. 5. 14. C-V log-log plot for Cu concentrations, excluding the three isolated boreholes

Table. 5. 9. Cu thresholds defined by the C-V model in the Kahang deposit, ignoring the three
isolated boreholes

Geochemical population Cu (wt.%) threshold value Range Cu (wt.%)

First (Barren host rock) - <0.071

Second (Main mineralisation
0.071 0.071-0.40
starting)
Third 0.40 0.40-1.86

Fourth 1.86 (Enriched zone for Cu) >1.86

219
The Cu average values with and without the three boreholes are 0.164 wt.% and 0.166

wt.% respectively, and the Cu histograms are similar in the both scenarios. Moreover, the

Cu estimated histogram in the new scenario (Fig. 5. 15) is similar to the Cu estimated

from all of the bore holes (See chapter 3). The comparison between two block models is

depicted in Table. 5. 10.

Fig. 5. 15. Estimated Cu histogram, ignoring the 3 boreholes located in the NW part of the
deposit

220
 Table. 5. 10. Comparison between results obtained from the two Cu block models

Average Grade Average Total Voxel Cu Range for

Cu Block Model Ore Tonnage
(wt.%), Raw Grade (wt.%), Amount Enriched zone
Data Estimated (wt.%)

48 Boreholes 210,080,697 0.164 0.14 489,927 1.86-3.24

45 Boreholes by
Ignoring the Three 0.152 263,410 >1.86
112,950,208 0.166
Isolated Boreholes
Located in the NW

5.12 Determination of an Ultimate Pit Limit when Ignoring the Three

Isolated Boreholes

The pit located with an elevation equal to 2210 m (Table. 5. 11) is determined as the new

Kahang final pit limit in which the NPV value corresponds to $3,731,732. Furthermore,

the ore and waste, strip ratio and mining lifetime report values of 1,475,582 tonnes,

4,713,207 tonnes, 3.19 and 1.48 years, respectively (Table. 5. 12).

Table. 5. 11. Pit limit reserve of the comparative case study

221
 Table. 5. 12. Ultimate pit characteristic for the comparative case study

Cumulative Data
Processing Mining Strip
Profit Revenue NPV Total Rock Total Ore Total Waste
Cost Cost Ratio
$ $ $ $ $ Tonnes Tonnes Tonnes
4,301,274 21,355,959 6,494,525 10,560,160 3,731,732 6,188,790 1,475,582 4,713,207 3.194

The total number of nested pits existing in this scenario reaches 100 extraction

sequences and five pit shells (Table. 5. 13). The sequence No. 90 is determined as the

OES with NPV equal to $3,590,462. In addition, the total ore and waste and

corresponding strip ratio at this point reports as 1,349,158 tonnes, 4,228,982 tonnes and

3.13, respectively (Fig. 5. 16).

222
 Fig. 5. 16. Cumulative NPV and profit values for the comparative case study driven by NPV
Scheduler (the black arrow indicates sequence No. 90 which specifies the Kahang ultimate pit
limit)

Table. 5. 13. Pit optimisation phases for the comparative case study

223
5.13 Results

One of the most crucial parameters in mining is the cut-off grade (COG) which defines

the grade for discriminating between ore and waste in an orebody over a mine life. The

results of this study show that the economic COG of the Kahang for Cu is 0.07 wt.% which

is almost equal to the commencement of Cu mineralisation resulted by the C-V fractal

model (See chapter 3, section 3.7.2).

Determination and analysis of the complete dataset, and the run without the three remote

drill holes, indicates that this deposit shows a positive NPV meaning that it is financially

feasible to produce Cu (as the main target) and Mo (by-product) for decision makers. The

pits located at elevations equal to 2225 m and 2210 m were determined as the Kahang

final pit limit for the complete dataset and dataset without the remote drill holes,

respectively (Tables. 5. 7 and 5. 11). The differences between the two scenarios

mentioned above in terms of the ultimate pit characteristics are shown in Table. 5. 14.

Table. 5. 14. Differences between ultimate pit limits characteristics of the complete dataset and
without the three remote drill holes

Ore Waste Mine Lifetime

NPV ($) Strip Ratio
(Tonne) (Tonne) (Year)

Complete Dataset 3,032,862 3,648,294 13,970,954 3.8 3.65

Dataset without
the Three Remote 3,731,732 1,475,582 4,713,207 3.19 1.48
Drill Holes

224
The NPV value of the pit without the three isolated drill holes is higher than the pit

generated with the complete dataset, although the productivity of the first scenario is

higher than the pit without three remote drill holes.

From a comparison of the two ultimate pit limit scenarios the sequence No. 92 was

determined as the Kahang optimal extraction sequence (OES) with respect to the

complete dataset. However, for the run without the three remote drill holes, sequence No.

90 was identified as the OES (Table. 5. 15).

Table. 5. 15. Differences between optimal extraction sequences characteristics of the complete
dataset (sequence No. 92) and without the three remote drill holes (sequence No. 90)

Ore Waste Cumulative

NPV ($) Strip Ratio
(Tonne) (Tonne) Profit Value
($)
Complete Dataset 2,884,968 3,291,944 12,901,028 3.91 7,853,825
Dataset without
the Three remote 3,590,462 1,349,158 4,228,982 3.13 4,128,521
Drill Holes

From the optimisation models run for the two scenarios, the cumulative profit value for

the Kahang deposit is lower when ignoring the three remote drill holes which may be due

to lower productivity (e.g., ore tonnage). From this, completion of more comprehensive

and systematic drilling in the deposit, especially to overcome the gap between boreholes

shown in Fig. 5. 11, seems sensible as it will likely increase the reserve due to better

geological constraints.

225
 CHAPTER SIX. Present Value-Volume (PV-V)
Fractal Modelling for Mining Strategy Selection

226
6.1 Introduction

The definition of optimal pit limits and profit is a fundamental part of prefeasibility and

feasibility studies in open pit mines (Koenigsberg, 1982; Dowd and Onur, 1993; Whittle,

1998b; Bernabe, 2001; Dincer, 2001; Dimitrakopoulos et al., 2007; Osanloo et al., 2008a;

Armstrong and Galli, 2012). The pit limit defines the ore and waste tonnages and the ore

values. The OES of the optimal pit represents a maximised Net Present Value (NPV).

The classical problem of pit optimisation is solvable using well-known and efficient

algorithms like the Lerch and Grossmann (1965) in order to reach the highest value of

DCF (Picard, 1976; Bond, 1995; Seymour, 1995; Hustrulid and Kuchta, 2006; Yasrebi et

al., 2011; Mart and Markey, 2013). In practice, pit optimisation is performed on voxels

whose true grades are unknown and can only be estimated or simulated using the

available information. However, future metals’ commodity prices are uncertain (Dowd,

1994; Marcotte and Caron, 2013). One of the key pieces of information required is an

optimal determination of the COG which depends on all of the salient technological

features of mining, such as the capacity of extraction and of milling, the geometry and

geology of the orebody and the optimal grade of ore to send for processing (Dagdelen

and Mohammed, 1997; Cairns and Shinkuma, 2003). Following this, Krautkraemer (1988)

found that the COG changing rate depends on the difference between the price and the

rate of interest. As a result, an increase in the metal’s commodity price reduces the COG

(Cairns and Shinkuma, 2003). Alternatively, when the metal price drops producers

attempt to mine ores with higher grades. Consequently, it is often necessary to design a

mining scenario (excavation orientation) to optimise pay-back in order to overcome the

problems of unpredictability of commodity price and variable mining expenses because

227
of maturity time, geological factors, engineering parameters, economic conditions and

political issues which can all influence the economic regime (Costa Lima and Suslick,

2006; King, 2011).

The project value is a linear function with respect to commodity price (e.g., Costa Lima

and Suslick, 2006; Asad and Dimitrakopoulos, 2013; Fig. 6. 1). However, in real cases,

project value is a nonlinear function due to the effects of grade distributions, fixed and

variable costs due to the spatial location of mineable voxels (Dimitrakopoulos et al., 2002;

Costa Lima and Suslick, 2006; Elkington and Durham, 2011; Marcotte and Caron, 2013).

Open pit mine design and long-term sequencing is an intricate and critically important part

of mining ventures. It provides the technical plan to be followed from mine development

to mine closure which has a profound effect on the economic value of the mine. Therefore,

the aim of this chapter is to propose a Present Value-Volume (PV-V) fractal model to

identify an accurate excavation orientation with respect to the economic principals of all

voxels within the Cu-Mo block model. This is obtained using the C-V fractal model and

voxels located within the determined ultimate pit limit which will take into account best

mining strategy.

228
 Fig. 6. 1. Linear relationship between the NPV and metal prices

6.2 Methodology

Based on the Cu-Mo block model of the Kahang deposit obtained by means of the C-V

fractal model (See chapter 3), a dataset was created for economic modelling including

the voxel’s coordination, density (See chapter 4), Cu and Mo values and metal prices,

rock type, mining and processing costs, recoveries and revenue (economic principals)

with respect to the each voxel. The PV values were calculated for each voxel.

Subsequently, a PV-V fractal model was generated for classification of the voxels’ values

in terms of profitability regarding positive values, as depicted in Fig. 6. 2. The proposed

PV–V fractal model can be expressed as follow:

V(ρPV≤υ) ∞ ρNPV−a1; V(ρPV≥υ) ∞ ρPV–a2

Equation 6-1

229
Where V(ρPV≤υ) and V(ρPV≥υ) denote volumes (V) with PV values (ρPV) that are,

respectively, smaller and greater than PV threshold values υ. a1 and a2 are characteristic

exponents as fractal dimensions.

Fig. 6. 2. Present Value (PV) block model for the Kahang Cu-Mo porphyry deposit (the grey
platform distinguishes the boundary between open pit and underground mining surfaces based
on chapter 5, section 5.10.1 and Table. 5. 7)

Secondly, the excavation orientation (pushbacks) were defined based on the results

obtained from the fractal modelling. In addition, a NPV-Cumulative Total Ore (NPV-CTO)

fractal model, in line with mining sequences (Nested pits: See Appendix. H), was

proposed in order to validate an optimal extraction sequences (OES). The fractal model

is expressed in the following form:

230
 CTO(ρNPV≤υ) ∞ ρNPV−a1; CTO(ρNPV≥υ) ∞ ρNPV–a2

Equation 6-2

Where CTO(ρNPV≤υ) and CTO(ρNPV≥υ) reveal cumulative total ore (CTO) with NPV values

(ρNPV) that are, respectively, smaller and greater than NPV threshold values. υ defines

those CTO and a1 and a2 are characteristic exponents as fractal dimensions. For

calculation of CTO(ρNPV≤υ) and CTO(ρNPV≥υ), mining sequences with their corresponding

NPV values were used.

6.3 Statistical Characteristics

According to the PV calculation for each voxel of the Kahang block model, 86,650 voxels

consisting of the positive PV values were used. The PV histogram generated based on

its positive values (Fig. 6. 3) shows a PV mean equal to $2,933. Furthermore, a PV

median was found to be $1,668 which reveals that the majority of voxels with positive

PVs have values lower than the mean. Moreover, most of the voxels contain PV values

lower than $10,000 and also a few voxels (1,253 voxels) have a PV value higher than

$20,000. Finally, a histogram of cumulative NPV was generated based on the mining

sequences derived via the optimisation operation considering the positive NPV values,

as depicted in Fig. 6. 3. Twelve out of 100 nested pits have positive cumulative NPV

values (See Appendix. H). There is a multimodal distribution for this variable. The main

population has high values of cumulative NPVs which are greater than its mean

($2,370,497).

231
 (a)

(b)

Fig. 6. 3. a) Profit value histogram based on PV block model, and b) NPV histogram based on

the mining sequences for the Kahang deposit

232
6.4 Application of PV-V Model

According to the 3D PV block model, a PV-V fractal model has been created which, from

log-log plot, shows that there are four populations corresponding to $501, $1,995,

$19,054 and $31,623 (Fig. 6. 4). This indicates that there is a multifractal nature in terms

of the PV values within the deposit. The voxels with high and extreme values of PV

commence from $19,054 and $31,623, respectively, which exist in the NE and central

parts of the area (Fig. 6. 5). The majority of voxels have PV values between $1,995 and

$19,054 in the deposit (third population in the log-log plot) entitled moderate population

of PV, as depicted in Fig. 6. 5. The number of voxels is around 112,000 and 134,000 in

the weak and moderate populations based on the PV-V model, as depicted in Table. 6.

1. Low amounts of voxels (<2600) contain the high and extreme populations with high

values of PV. Most of the voxels with high and extreme PV values exist at depth, in the

NW and especially in the central parts, as depicted in Fig. 6. 5.

Fig. 6. 4. PV log-log plot

233
(a)

(b)

234
 (c)
Fig. 6. 5. PV distribution within the deposit based on the PV-V fractal modelling consisting of a)

moderate population, b) high PVs, and c) extreme population

Table. 6. 1. PV thresholds defined by PV-V model in the Kahang deposit

PV Population PV ($) Range Number of Voxels

Very weak <501 57305
Weak 501-1995 111994
Moderate 1995-19054 133757
High 19054-31623 1925
Extreme ≥31623 604

Based on the ultimate pit limit (See chapter 5), most of the voxels with high and extreme

PV values are situated in the deeper parts of the mine, as depicted in Fig. 6. 6. The

majority of these voxels are close to the open pit limit, especially in the central part, which

235
means that an ultimate pit limit can be moved deeper if detailed grid drilling can be carried

out. The PV plans in the different excavation levels were derived via a PV block model

and classified on the basis of PV-V fractal modelling to propose an accurate mining

orientation in order to achieve an earlier pay-back period (Fig. 6. 7). As a result, the PV

values increase from the north of the deposit to the south which is defined as the

excavation orientation.

Fig. 6. 6. Voxels with high and extreme PV values within the deposit (the grey platform
distinguishes the boundary between open pit and underground mining surfaces based on
chapter 5, section 5.10.1 and Table. 5. 7)

236
(a)

237
(b)

238
(c)

239
(d)

240
(e)

241
(f)

242
(g)

243
 (h)
Fig. 6. 7. PV plan views based on the PV-V fractal model in elevations of: a) 2230 m, b) 2240 m,

c) 2250 m, d) 2260 m, e) 2270 m, f) 2280 m, g) 2290 m, and h) 2300 m

244
6.5 Application of NPV-CTO Model

The criteria used to determine an OES is subjective in the sense that the chosen

sequence is the one that has the NPV nearest to the highest NPV calculated via NPV

Scheduler. The proposed NPV-CTO fractal model is a reward of developing a new

method to reduce the number of required mining sequences and practical time to obtain

an optimal solution to determine an OES, especially in an absence of computer-based

optimisation software (i.e., NPV Scheduler). In addition, this model reduces the number

of analyses and data transfer processes that are often necessary in standard computer-

based practice for open pit optimisation. This is necessary to overcome probable mining

risks due to uncertainty resulting from e.g. any sudden economic changes or a decrease

in the commodity metal price especially if this happens at the end of mine life (Godoy and

Dimitrakopoulos, 2004; Montiel and Dimitrakopoulos, 2013). However, the best selection

of mining sequence is determined when the NPV cumulative trend becomes steady (See

chapter 5, Fig. 5. 10). This selection for identification of an OES is controlled manually,

typically from a nested pit shell methodology based on the experience of the engineer,

and consequently an optimum solution for this problematic issue cannot be developed

and it may lead to suboptimal results (Lerch and Grossmann, 1965; Bond, 1995; Hustrulid

and Kuchta, 2006; Elkington and Durham, 2011; Mart and Markey, 2013). As a result, a

mathematical method to provide an analytical practice, which intends to prevent manual

identification of an OES, seems to be inevitable.

In the author’s view, a proposed model should be rigorously tested against those already

available, and the possible errors discussed. Therefore, results of the proposed fractal

245
model were compared with the results of the OES, generated from NPV Scheduler (See

chapter 5 and Appendix. H). The NPV-CTO log-log plot indicates a mono-fractal nature

meaning that there is only one threshold value which corresponds to $2,754,229 and

3,288,516 tonnes of minable ore (Fig. 6. 8). The result of an obtained OES (from Chapter

5), which is the point that mining will be suspended due to a steady trend in the cumulative

NPV, is close to the result achieved through the NPV-CTO model. Possible errors for this

are shown in Table. 6. 2.

Table. 6. 2. Comparison between OESs regarding cumulative NPV and ore, calculated from
NPV Scheduler and NPV-CTO fractal model

Identification of an Optimal Extraction Sequence (OES)

Extraction Sequence Chart via Extraction Sequence Chart via Errors

NPV Scheduler NPV-CTO model

NPV ($) Ore (Tonne) NPV ($) Ore (Tonne) NPV (%) Ore (%)

2,884,968 3,291,944 2,754,229 3,288,516 4 0.1

246
 Fig. 6. 8. The NPV-CTO log-log plot in the Kahang deposit

6.6 Results

From the PV-V log–log plot from Kahang, there is a multifractal PV distribution within the

deposit. In addition, there are five threshold values meaning four PV populations for the

deposit. The threshold value of $1995 is the commencement of the moderate population

which occurs in the majority of voxels within the deposit. The high PV values are present

in the NE and central parts of the deposit. The extreme values are situated at the depth

and NW parts, around the three remote drill holes. The 2D and 3D maps for PV

distribution show that the profitability increases from the north to south of the deposit

proposing an open pit excavation orientation to achieve an earlier pay-back.

247
The NPV-Cumulative Total Ore (NPV-CTO) fractal model was proposed in order to find

an OES. The NPV-CTO log–log plot from the Kahang deposit shows that there is a mono-

fractal model which has a threshold value for cumulative NPV and enclosing ore equal to

$2,754,229 and 3,288,516 tonnes, respectively. This indicates that the mining operation

can be terminated when reaching this point to reduce the number of required mining

sequences. This overcomes probable mining risks due to uncertainty of the relative

metal’s world commodity price and sudden mining costs.

The comparison between the results for OES calculated from NPV Scheduler and NPV-

CTO fractal model shows that the errors for NPV and minable ore are 4% and 0.1%,

respectively. This may suggest that the developed model is reliable and can be used

beyond the Kahang deposit in determination of an OES for open pit mines.

248
CHAPTER SEVEN. Conclusions and
Recommendation for Future Work

249
Conventional methods including calculation of mean and standard derivation (SD),

probability graphs, explorational data analysis (EDA) and multivariate data analysis have

been widely used in geochemical exploration. However, these methods do not consider

spatial variations in geochemical patterns. In the past decades, a number of complex

structures and phenomena have been quantitatively characterised by fractal/multifractal

modelling (the most commonly used fractal models have been introduced in this thesis).

The utility of fractal/multifractal models for geochemical data is to delineate geochemical

populations and quantify the spatial distribution of geochemical data. A variety of

fractal/multifractal models for this purpose have been introduced and used in different

kinds of deposits (e.g., Qulong copper deposit, Tibet, western China; Sungun porphyry

copper deposit, Iran; Mitchell Sulphurets precious metal district, British Columbia,

Canada; Cambrian Hellyer volcanic-hosted massive sulphide deposit, Australia;

undiscovered mineral deposits in Gejiu, Yunnan province, China; Zaghia iron ore deposit,

central Iran; Tangedezan Pb–Zn carbonate hosted deposits, central Iran; bauxite

orebodies in the Guangxi province, China). The fractal/multifractal modelling has been

shown to be a useful tool for mineral exploration, rock mechanics and economical

evaluation of the Kahang porphyry deposit due to ore elements, rock mass and

economical parameter variation. The advantages of the fractal modelling, using C-V, D-

V, RQD-V, PV-V and NPV-CTO, is its simplicity and easy computational implementation,

as well as the possibility to compute numerical values for variables, e.g., geochemical

data, density, RQD and present value (PV) thresholds, which are deemed to be the most

useful criteria for cross examination of data.

250
Evaluation of ore element distribution via the C-V fractal model was carried out to

separate ‘barren’ host rock, below the cut-off grade in an open pit optimisation, from

mineralised zones, especially for the cases in which element concentrations occur in the

various geological zones. Such complexity can be efficiently recognised by

fractal/multifractal analysis such as C-V fractal model using log–log plots.

A comparison between the resulting 3D models from multifractal analysis and traditional

statistical methods indicates that the statistical methods can only consider the elemental

concentration and they ignore the spatial variability in the block models of the deposits

which may appear within the model. In fractal models, the spatial correlation of data is of

interest such as Cu and Mo grades, density, RQD and PV. Statistical analysis applied to

the data has shown non-normal distribution for Cu, Mo, density, RQD and PV.

Accordingly, in statistics, only one threshold can be extracted for each element which is

the mean value. The multifractal model provides several thresholds separating various

stages of the regional variables.

The Kahang Cu-Mo porphyry deposit in this case study consists of Eocene volcanic–

pyroclastic rocks which were intruded by Oligo-Miocene porphyric granitoids rocks, quartz

monzonite, monzodiorite-monzogranites and diorites. The geological results from

lithology, alteration and zonation and also subsurface geochemical data including Cu and

Mo values in this study have shown a porphyry deposit in this area. The main host rock

is porphyritic quartz diorite for Cu-Mo mineralisation. The alteration map in Fig. 2. 5a

shows four major types of hydrothermal alterations in the Kahang deposit: potassic,

251
phyllic, argillic and propylitic. The most extensive hydrothermal alteration zone in the

Kahang is phyllic. Minor amounts of chalcopyrite and molybdenite are seen in this zone,

but the major sulphide mineral is pyrite. The Cu mean value within this zone is equal to

0.14 wt.% based on raw data from drill holes. This zone has occupied a large part of the

deposit from depth to surface (Fig. 2. 13). Potassic alteration is observable in the NW and

deep parts of the deposit. The common potassic alteration zone in Kahang was

distinguished from the presence of mineral assemblages (e.g., KF, secondary biotite,

quartz veins and veinlets, magnetite and chlorite). Moreover, there is evidence of potassic

alteration to the west, especially around the three remote drill holes located in the NW

part, which can suggest that mineralisation within this zone may continue to the west (See

Cu distribution with different thresholds in the NW drillholes in Fig. 3. 20). In addition, the

Cu mean value within the potassic zone is 0.12 wt.% based on the raw data. Argillic

alteration is seen on surface. Main products of this alteration is kaolinite which was

produced from alteration of plagioclase phenocrysts and groundmass. Jarosite, as the

second major alteration mineral in this zone is present (Fig. 2. 8c). The propylitic alteration

zone is developed in marginal parts of the deposit. The most important products of this

alteration in order of abundance are chlorite, calcite and minor epidote (Fig. 2. 9). Based

on abundance of chlorite and calcite, this alteration zone is divided into two main chloritic

and calcitic parts. Hypogene-type mineralisation hosts most Cu ore in the Eastern part of

the Kahang deposit but the supergene enrichment zone is relatively small and occurs in

the central part of the area which approaches the surface.

252
According to the C-V fractal model, the main threshold values for Cu and Mo are 0.42

wt.% and 100 ppm, respectively. Enriched Cu-Mo mineralised zones with Cu ≥ 1.8 wt.%

and Mo ≥645 ppm are located in the central, NW and NE parts within the hypogene zone.

The supergene enrichment zone occurs in small areas within the deposit, especially in

the central and eastern parts close to the surface. The hypogene and supergene

enrichment zones outlined by the C-V model correlate well with the alterations and

mineralogical data shown in the 3D geological models. The C-V log-log plots from the

Kahang deposit show that there is a multifractal model for Cu and Mo. Correlation

between the results of the C-V model and the chosen geological particulars show that the

supergene enrichment zone has a high correlation within the chalcocite accumulations

within the Kahang deposit. The main hypogene zone has an association with the

chalcopyrite distribution model having Cu ≥ 0.42 wt.%. According to the correlation

between results driven by fractal modelling and geological models by a logratio matrix,

the main Cu and Mo mineralised zones generated by the C-V fractal model have a strong

correlation with the potassic alteration zone with respect to the overall accuracy.

In this research, the D–V (Density–Volume) fractal model has been proposed to delineate

rock mass characteristics. The results from the D–V fractal model have been correlated

with the major rock types and validated against an RQD model. The D–V model has been

successfully applied to model relationships between density values and volumes in the

Kahang Cu–Mo porphyry deposit. The D–V log–log plot from the Kahang deposit

indicates that there is a mono-fractal model which has a breakpoint in density of 2.7t/m3.

253
Correlation between the results of D–V and RQD models reveals that the rock units with

a higher density (>2.7t/m3) are associated with RQD values > 70%. The final pit slope

geometry and ultimate pit limit will depend on the economic evaluation of the ore body.

An awareness of the spatial variability of parameters such as RQD and density can be

used to assess the geotechnical characteristics of the rock mass. This can then be used

to evaluate potential slope stability and be incorporated into a geotechnical risk model for

the final pit geometry. Regions of high RQD may be targeted as offering greater potential

for increased slope angles or locations for siting of critical haul roads. Regions of lower

RQD should, where possible, be avoided for final pit limits. Further geotechnical

characterisation will, however, be necessary to establish any potential influence of the 3D

fracture network and presence of any major discontinuity-controlled instability.

The RQD-Volume (RQD-V) fractal model was used to investigate and delineate various

RQD populations in the Kahang Cu-Mo porphyry deposit (Central Iran). The RQD-V

fractal model illustrates four RQD populations in the deposit. The RQD threshold value

for excellent rocks is 89.1% based on the fractal model as situated in the central and NW

parts of the deposit. Models of good and fair rocks in the central, eastern and NW parts

of the deposit contain 25.1–89.1% RQD according to the RQD-V model. Furthermore, the

correlation between results driven by the fractal modelling and major lithological unit

(PQD) in the Kahang deposit, rocks with excellent RQD defined by the means of the RQD-

V model have a strong correlation with porphyritic quartz diorite shown by the 3D

geological model.

254
The numbers of mining sequences (nested pits) were determined during optimisation

studies using an exported dataset of the Cu-Mo block model shown by the means of the

C-V fractal model. Furthermore, various combinations of nested pits (pit No.1 to pit No.

100) have been discussed and finally pit No. 92 was selected as the closure of the open

pit mining of the Kahang deposit.

The need for copper especially in developing countries like Iran is of paramount

importance therefore, the first choice of Cu block model considering all 48 boreholes is

recommended due to the higher level of productivity in comparison to the second

scenario, however the NPV of the second scenario is higher than the NPV of the deposit

including all 48 boreholes. Therefore, a minable reserve exists in the pit consisting of all

completed boreholes of 3,648,294 tonnes of sulphide ore which indicates the range of

greater productivity compared to the final pit, ignoring the three isolated boreholes in the

NW section of the studied deposit.

With respect to the C-V fractal model histograms, there is no major difference in terms of

Cu average grades for either of the Cu block models however; there are remarkable

differences between ore tonnage and the total voxel count considering the two block

models as depicted in Table. 5. 10.

The PV-V and NPV-CTO fractal models have been proposed to delineate economic

parameters. The voxels’ values were classified according to the PV-V fractal model in the

deposit which reveal that the open pit limit can be deeper if grid drilling can be developed.

255
Supporting that, the deposit sections in terms of elemental concentrations (Cu as the main

target) is low within the initial years of excavation and it will be gradually upgraded at

depth especially at plan level 2225 m in terms of Z direction (elevation) as depicted in the

Fig. G. 2 of Appendix. G and Fig. 6. 6 of the Chapter 6. Consequently, the results obtained

by the PV-V fractal model show that the majority of the orebody can be extracted by an

underground mining operation. The threshold values obtained by the means of the PV-V

fractal modelling suggest that the PV values have an increase from the north of the

deposit to the south which can be used as an indicator for determination of the excavation

orientation.

Open pit mine design and determination of mining orientation are a critically important

part of a mining venture from mine development to mine closure and have a profound

effect on the economic value of the mine. The most established and frequently employed

practice to mine closure since the 1980s is based on the Lerch and Grossman three

dimensional graph theory to determine the best mining sequence by which the mining

operation (ore and waste excavation within the pit) will be terminated. However, the

proposed NPV-CTO fractal model provided an analytical tool which can be used for

determination of an OES for an open pit mine. The OES results via NPV Scheduler is

much the same with the results obtained by the suggested model (Table. 6. 2) meaning

that the NPV-CTO fractal can be implemented in the absence of optimisation software

packages.

256
Since hydrological and geotechnical studies have not been undertaken in the study area,

one is really forced to do such tests and experiments to calculate and evaluate the total

gradient of the open pit mine’s slope walls.

The necessity of reviewing the cut-off grade throughout the project lifetime and specifying

the grade limit using optimisation models to increase the project NPV is highly

recommended simultaneously with variations in the world metal commodity price.

Furthermore, the general geological observation and careful consideration of geological

features of the Kahang deposit suggests that the mineralisation continues to the west and

east and even also towards the Kahang village. As such a more comprehensive and

systematic drilling programme is recommended in order to better characterise the Kahang

deposit, which may correspondingly increase its estimated resource.

Mineral resource classification is important in uncertainty assessment and risk analysis.

The Joint Ore Reserves Committee (JORC) is widely utilised for this purpose which

classifies mineral resources as measured, indicated and inferred, depending on the

degrees of confidence. Ore reserves are classified as proven and probable from either

measured or indicated mineral resource. However; it is recommended for those who are

interested in the area of this research to establish an innovative application based on the

combination of geostatistical simulation (e.g., turning bands simulation) and fractal

modelling for mineral resource classification.

257
An accurate description for geological domains which discriminates types of mineralogy,

alteration and lithology is an important task in mineral resource and reserve evaluation.

Deterministic models, based on drill hole data, define just one layout of these domains

and do not consider uncertainty in a study area so they cannot measure uncertainty in

the domain boundaries. However, stochastic models with respect to geostatistical

simulation (especially plurigaussian) have distinctive power to assess uncertainty in the

spatial layout of the domains which contribute to enhanced geological control for the

quantitative variables of interest (e.g., porosity, permeability and concentration).

Therefore, the use of a plurigaussian simulation is recommended to determine geological

domains that control the grade distribution to obtain a final grade model for the Kahang

deposit.

258
References

259
Afshooni, S.Z., Asadi Harooni, H., Esmaili, D., 2011. The microthermometry study of fluid

inclusions in quartz veins of Kahang deposit (north eastern of Isfahan). 2 nd National

Symposium of Iranian Society of Economic Geology. Lorestan University, p. 144 (In

Persian with English abstract).

Afshooni, S.Z., Esmaili, D., Asadi Harooni, H., 2010. The study of mineralography in

Kahang exploration area (NE Isfahan). Geophysical Research Abstracts: EGU

General Assembly, 13, p. 318.

Afshooni, S.Z., Mirnejad, H., Esmaeily, D., Asadi Haroni, H., 2013. Mineral chemistry of

hydrothermal biotite from the Kahang porphyry copper deposit (NE Isfahan), Central

Province of Iran. Ore Geology Reviews 54, 214–232.

Afzal, P., Ahari, H.D., Omran, N.R., Aliyari, F., 2013c. Delineation of gold mineralized

zones using concentration–volume fractal model in Qolqoleh gold deposit, NW Iran.

Ore Geology Reviews 55, 125–133.

Afzal, P., Alhoseini, S.H., Tokhmechi, B., Kaveh Ahangarana, D., Yasrebi A.B., Madani,

N., Wetherelt, A., 2014. Outlining of high quality coking coal by concentration–

volume fractal model and turning bands simulation in East-Parvadeh coal deposit,

Central Iran. International Journal of Coal Geology 127, 88-99.

Afzal, P., Dadashzadeh Ahari, H., Rashidnejad Omran, N., Aliyari, F., 2013a. Delineation

of gold mineralized zones using concentration–volume fractal model in Qolqoleh

gold deposit, NW Iran. Ore Geology Reviews 55, pp. 125–133.

Afzal, P., Fadakar Alghalandis, Y., Khakzad, A., Moarefvand, P., Rashidnejad Omran, N.,

2011. Delineation of mineralization zones in porphyry Cu deposits by fractal

260
 concentration–volume modeling. Journal of Geochemical Exploration 108, 220–

232.

Afzal, P., Fadakar Alghalandis, Y., Moarefvand, P., Rashidnejad Omran, N., Asadi

Haroni, H., 2012. Application of power-spectrum–volume fractal method for

detecting hypogene, supergene enrichment, leached and barren zones in Kahang

Cu porphyry deposit, Central Iran. Journal of Geochemical Exploration 112, 131–

138.

Afzal, P., Harati, H., Fadakar Alghalandis, Y., Yasrebi, A.B., 2013b. Application of

spectrum–area fractal model to identify of geochemical anomalies based on soil

data in Kahang porphyry-type Cu deposit, Iran. Chemie der Erde - Geochemistry

739 (4), 533-543.

Afzal, P., Khakzad, A., Moarefvand, P., Rashidnejad Omran, N., Esfandiari, B., Fadakar

Alghalandis, Y., 2010. Geochemical anomaly separation by multifractal modeling in

Kahang (Gor Gor) porphyry system. Central Iran. Journal of Geochemical

Exploration 104, 34–46.

Agard, P., Omrani, J., Jolivet, L., Mouthereau, F., 2005. Convergence history across

Zagros, Iran: constraints from collisional and earlier deformation. International

Journal of Earth Sciences 94, 401–419.

Aghanabati, A., 2004. Geology of Iran. Geological Survey of Iran, 587 pp (In Persian).

Aghazadeh, M., Hou, Z., Badrzadeh, Z., Zhou, L., 2015. Temporal–spatial distribution and

tectonic setting of porphyry copper deposits in Iran: Constraints from zircon U–Pb

and Molybdenite Re–Os geochronology. Ore Geology Reviews 70, 385–406.

261
Agterberg, F.P., 2007. Mixtures of multiplicative cascade models in geochemistry.

Nonlinear Processes in Geophysics 14, 201–209.

Agterberg, F.P., 2012. Multifractals and geostatistics. Journal of Geochemical Exploration

122, 113–122.

Agterberg, F.P., Cheng, Q., Brown, A., Good, D., 1996. Multifractal modeling of fractures

in the Lac du Bonnet batholith, Manitoba. Computers and Geosciences 22 (5), 497–

507.

Agterberg, F.P., Cheng, Q., Wright, D.F., 1993. Fractal modeling of mineral deposits. In:

Elbrond, J., Tang, X. (Eds.), 24th APCOM symposium proceeding, Montreal,

Canada, pp. 43–53.

Ahmadian, J., Haschke, M., McDonald, I., Regelous, M., Ghorbani, M., Emami, M.,

Murata, M., 2009. High magmatic flux during Alpine–Himalayan collision: constraints

from the Kal-e-Kafi complex, central Iran. Geological Society of America Bulletin

121, 857–868.

Ahrens, L.H., 1954. The lognormal distribution of elements (a fundamental law of

geochemistry and its subsidiary). Geochimica et Cosmochimica Acta 5, 49–73.

Ahrens, L.H., 1957. Lognormal-type distributions-III. Geochimica et Cosmochimica Acta

11, 205–212.

Akaike A., Dagdelen K., 1999. A strategic production scheduling method for an open pit

mine. In: Proceedings of the 28th APCOM, Golden Colorado, USA, pp. 729-738.

Akbari, A.D., Osanloo, M., Shirazi, M.A., 2008. Determination of ultimate pit Limits in open

Mines Using Real Option Approach. IUST International Journal of Engineering

Science 19, 23–38.

262
Alavi, M., 1994. Tectonic of Zagros orogenic belt of Iran: new data and interpretations.

Tectonophysics 229, 211-238.

Alavi, M., 2004. Regional stratigraphy of the Zagros folded-thrust belt of Iran and its

proforeland evolution. American Journal of Science 304, 1-20.

Albanese, S., De Vivo, B., Lima, A., Cicchella, D., 2007. Geochemical background and

baseline values of toxic elements in stream sediments of Campania region (Italy).

Journal of Geochemical Exploration 93, 21–34.

Ali, K., Cheng, Q., Zhijun, C., 2007. Multifractal power spectrum and singularity analysis

for modelling stream sediment geochemical distribution patterns to identify

anomalies related to gold mineralization in Yunnan Province, South China.

Geochemistry: Exploration, Environment, Analysis 7 (4), 293–301.

Alpers, C.N., Brimhall, G.H., 1989. Paleohydrologic evolution and geochemical dynamics

of cumulative supergene metal enrichment at La Escondida, Atacama Desert,

Northern Chile. Economic Geology 84, 229–255.

Amit, K., Chatterjee, D., Jacks, G., 2014. Shallow hydrostratigraphy in an arsenic affected

region of Bengal Basin: Implication for targeting safe aquifers for drinking water

supply. Science of the Total Environment 485–486, 12–22.

Armstrong, D., 1990. Definition of Mining Parameters. In Surface Mining, 2 nd Edition (Ed.

B.A. Kennedy), Littleton: Society for Mining, Metallurgy and Exploration, Inc. pp.

459–464.

Armstrong, M., Boufassa, A., 1988. Comparing the robustness of ordinary kriging and

lognormal kriging: outlier resistance. Mathematical Geology 20 (4), 447–457.

263
Armstrong, M., Galli, A., 2012. New approach to flexible open pit optimisation and

scheduling. Mining Technology 121 (3), 132-138.

Asad, M.W.A., Dimitrakopoulos, R., 2013. Implementing a parametric maximum flow

algorithm for optimal open pit mine design under uncertain supply and demand.

Journal of the Operational Research Society 64, 185-197.

Asadi, H., Kianpouryan, S., Lu, Y., McCuaig, T.C., 2014. Exploratory data analysis and

C-A fractal model applied in mapping multi-element soil anomalies for drilling: A

case study from the Sari Gunay epithermal gold deposit, NW Iran. Journal of

Geochemical Exploration 145, 233–241.

Asadi, H., Porwal, A., Fatehi, M., Kianpouryan, S., Lua, Y.J., 2015. Exploration feature

selection applied to hybrid data integration modeling: Targeting copper-gold

potential in central Iran. Ore Geology Reviews (In Press, Corrected Proof).

Asadi, S., Moore, F., Zarasvandi, A., 2014. Discriminating productive and barren porphyry

copper deposits in the southeastern part of the central Iranian volcano-plutonic belt,

Kerman region, Iran: A review. Earth-Science Reviews 138, 25–46.

Asghari, O., Hezarkhani, A., 2008. Applying discriminant analysis to separate the

alteration zones within the Sungun porphyry copper deposit. Asian Journal of

Applied Sciences 8 (24), 4472—4486.

Asghari, O., Hezarkhani, A., Soltani, F., 2009. The comparison of alteration zones in the

Sungun porphyry copper deposit, Iran (based on fluid inclusion studies). Acta

Geologica Polonica 59 (1), 93–109.

264
Asghari, O., Madani Esfahani, N., 2013. A new approach for the geological risk evaluation

of coal resources through a geostatistical simulation Case study: Parvadeh III coal

deposit. Arabian Journal of Geosciences 6, 957-970.

Askari-Nasab, H., Pourrahimian, Y., Ben-Awuah, E., Kalantari, S., 2011. Mixed integer

linear programming formulations for open pit production scheduling. Journal of

Mining Science 47, 338-359.

Ataei, M., Osanloo, M., 2003. Methods for Calculation of Optimal Cutoff Grades in

Complex Ore Deposits. Journal of Mining Science 39 (5), 499-507.

Atapour, H., Aftabi, A., 2007. The geochemistry of gossans associated with Sarcheshmeh

porphyry copper deposit, Rafsanjan, Kerman, Iran: Implications for exploration and

the environment. Journal of Geochemical Exploration 93 (1), 47–65.

Awadelseid, S.F., Xie, S., Bao, Z., Lei, L., Pan, F., Xu, D., Abdel Rahman, A.A., 2015.

Multifractal distribution of geochemical elements in Dexing porphyry copper deposit,

Jiangxi Province, China. Journal of Geochemical Exploration 149, 30–42.

Ayati, F., Yavuz, F., Noghreyan, M., Asadi Haroni, H., Yavuz, R., 2008. Chemical

characteristics and composition of hydrothermal biotite from the Dalli porphyry

copper prospect, Arak, central province of Iran. Mineralogy and Petrology 94, 107–

122.

Azadi, M., Mirmohammadi, M., Hezarkhani, A., 2014. Aspects of magmatic–hydrothermal

evolution of Kahang porphyry copper deposit, Central Iran. Arabian Journal of

Geosciences, DOI 10.1007/s12517-014-1528-2.

265
Baecher, G.B., Lanney, N.A., Einstein, H.H., 1977. Statistical description of rock

properties and samples. In: Proceedings of 18th US Symposium on Rock

Mechanics, Colorado, USA, pp. 5c1-5c8.

Bastante, F.G., Taboada, J., Alejano, L., Alonso, E., 2008. Optimization tools and

simulation methods for designing and evaluating a mining operation .Stochastic

Environmental Research and Risk Assessment 22, 727–735.

Beane, R.E., 1982. Hydrothermal alteration in silicate rocks. In: Titley, S.R. (Ed.),

Advances in Geology of the Porphyry Copper Deposits, Southwestern North

America. The University of Arizona Press, Tucson, pp. 117–137.

Beane, R.E., Titley, S.R., 1981. Porphyry copper deposits, Part II: Hydrothermal alteration

and mineralization. Economic Geology 75, 235-269.

Behrens, J.T., 1997. Principles and Procedures of Exploratory Data Analysis.

Psychological Methods 2, 131–160.

Berberian, M., King, G.C.P., 1981. Towards a palaeogeography and tectonic evolution of

Iran. Canadian Journal of Earth Sciences 18, 210-265.

Berberian, F., Muir, I.D., Pankhurst, R.J., Berberian, M., 1982. Late Cretaceous and Early

Miocene Andean-type plutonic activity in northern Makran and Central Iran. Journal

of the Geological Society of London 139, 605-614.

Berger, B., Ayuso, R., Wynn, J., Seal, R., 2008. Preliminary Model of Porphyry Copper

Deposits. U.S. Geological Survey Open-File Report 2008, Reston, Virginia, 20–55.

Bernabe, D., 2001. Comparative analysis of open pit mine scheduling techniques for

strategic mine planning of a copper mine in Southern Peru. M.Sc. Thesis, Colorado

School of Mines, Golden CO.

266
Bieniawski, Z.T., 1984. Rock Mechanics Design in Mining and Tunnelling, Balkema

publisher, Rotterdam, 272 pp.

Boadu, F.K., Long, LT., 1994. The fractal character of fracture spacing and RQD.

International Journal of Rock Mechanics and Mineral Science- Geomechanics

Abstracts 31, 127–134.

Bölviken, B., Stokke, P.R., Feder, J., Jössang, T., 1992. The fractal nature of geochemical

landscapes. Journal of Geochemical Exploration 43 (2), 91–109.

Bond, G., 1995. A mathematical analysis of the Lerchs and Grossmann algorithm and the

nested Lerch and Grossmann algorithm, Ph.D. Dissertation, Colorado School of

Mines, Golden, USA.

Bondy, J.A., Murty, U.S.R., 1976. Graph Theory with Applications. Macmillan, The

University of California.

Boomeri, M., Nakashima, K., Lentz, D.R., 2009. The Miduk porphyry Cu deposit, Kerman,

Iran: A geochemical analysis of the potassic zone including halogen element

systematics related to Cu mineralization processes. Journal of Geochemical

Exploration 103 (1), 17-19.

Caccetta, L., Giannini, L., 1986. Optimisation Techniques for the Open Pit Limit Problem.

In: Proceedings of Australian Institute of Mining and Metallurgy 291 (8), pp. 57-63.

Caccetta, L., Giannini, L.M., 1988. An application of discrete mathematics in the design

of an open pit mine. Discrete Applied Mathematics 21 (1), 1–19.

Cairns, R.D., Shinkuma, T., 2003. The choice of the cut-off grade in mining. Resources

Policy 29, 75–81.

267
Carranza, E.J.M., 2008. Geochemical anomaly and mineral prospectivity mapping in GIS.

Handbook of Exploration and Environmental Geochemistry, Vol. 11. Elsevier,

Amsterdam. 351 pp.

Carranza, E.J.M., 2009. Controls on mineral deposit occurrence inferred from analysis of

their spatial pattern and spatial association with geological features. Ore Geology

Reviews 35, 383–400.

Carranza, E.J.M., 2011. Analysis and mapping of geochemical anomalies using logratio-

transformed stream sediment data with censored values. Journal of Geochemical

Exploration 110, 167—185.

Carranza, E.J.M., Owusu, E., Hale, M., 2009. Mapping of prospectivity and estimation of

number of undiscovered prospects for lode-gold, south western Ashanti Belt,

Ghana. Mineralium Deposita 44 (8), 915–938.

Carranza, E.J.M., Sadeghi, M., 2010. Predictive mapping of prospectively and

quantitative estimation of undiscovered VMS deposits in Skellefte district (Sweden).

Ore Geology Reviews 38, 219–241.

Carrasco, P., Carrasco, P., Jara, E., 2004. The economic impact of correct sampling and

analysis practices in the copper mining industry. Chemometrics Intelligent

Laboratory Systems 74, 209–213.

Carvalho, J.L., Carter, T.G., Diederichs, M.S., 2007. An approach for prediction of

strength and post yield behaviour for rock masses of low intact strength. Rock

Mechanics: Meeting Society's Challenges and Demands, In: Proceedings of the 1st

Canada-US Rock Mechanics Symposium, Vancouver, Canada, pp. 277-285.

268
Cheng, Q., 1999. Spatial and scaling modelling for geochemical anomaly separation.

Journal of Geochemical Exploration 65 (3), 175–194.

Cheng, Q., 2007. Mapping singularities with stream sediment geochemical data for

prediction of undiscovered mineral deposits in Gejiu, Yunnan Province, China. Ore

Geology Review 32, 314–324.

Cheng, Q., 2012. Singularity theory and methods for mapping geochemical anomalies

caused by buried sources and for predicting undiscovered mineral deposits in

covered areas. Journal of Geochemical Exploration 122, 55–70.

Cheng, Q., Agterberg, F.P., 1996. Multifractal modeling and spatial statistics.

Mathematical Geology 28 (1), 1–16.

Cheng, Q., Agterberg, F.P., 2009. Singularity analysis of ore-mineral and toxic trace

elements in stream sediments. Computers and Geosciences 35 (2), 234–244.

Cheng, Q., Agterberg, F.P., Ballantyne, S.B., 1994. The separation of geochemical

anomalies from background by fractal methods. Journal of Geochemical Exploration

51, 109–130.

Clark, I., 1999. A case study in the application of geostatistics to lognormal and quasi-

lognormal problems. 28th International Symposium on Application of Computers and

Operations Research in the Mineral Industry. Colorado School of Mines, CSM, 407–

416.

Coghill, P., Miljak, D., Williams, E., 2014. Consequences of fractal grade distribution for

bulk sorting of a copper porphyry deposit. Geoscience Frontiers (In Press, Corrected

Proof).

269
Costa Lima, G.A., Suslick, S.B., 2006. Estimating the volatility of mining projects

considering price and operating cost uncertainties. Resources Policy 31, 86–94.

Cox, D., Singer, D., 1986. Mineral deposits models. U.S. geological survey bulletin. 1693

pp.

Cressie, N., 1993. Statistics for spatial data, John Wiley & Sons, New York, 900 pp.

Dagdelen, K., Mohammed, W.A.A., 1997. Multi mineral cutoff grade optimization with

option to stockpile. In: Proceedings of the SME Annual Meeting, Society for Mining,

Metallurgy and Exploration Inc., Littleton, CO, pp. 1–12.

Dargahi, S., Arvin, M., Pan, Y., Babaei, A., 2010. Petrogenesis of Post-Collisional A-type

granitoid from the Urumieh-Dokhtar magmatic assemblage, Southwestern Kerman,

Iran: Constraints on the Arabian- Eurasian continental collision. Lithos 115, 190-

204.

Darnley, A.G., Bjorklund, B., Gustavsson, N., Koval, P.V., Plant, J., Steenfelt, A., Tauchid,

T.M., Xie, X.J., 1995. A global geochemical database for environmental and

resource management. Recommendations for international geochemical mapping.

Earth Sciences Report 19. UNESCO Publishing, Paris.

Darwen, P.J., 2001. Genetic algorithms and risk assessment to maximize NPV with

robust open-pit scheduling. Fourth biennial conference on strategic mine planning,

pp. 29–34.

Das, M., Edgar, G.A., 2005. Separation properties for graph-directed self-similar fractals.

Topology and its Applications 152 (1–2), 138–156.

David. M., 1970. Geostatistical Ore Reserve Estimation, Elsevier, Amsterdam, 283 pp.

270
Davies, T.C., Mundalamo H.R., 2010. Environmental health impacts of dispersed

mineralisation in South Africa, Journal of African Earth Sciences 58, 652–666.

Davis. B., 1987. Uses and abuses of cross validation in geostatistics. Mathematical

Geology 19 (3), 241-48.

Davis. J.C., 2002. Statistics and Data Analysis in Geology, 3th ed. John Wiley & Sons Inc,

New York.

Deere, D.U., Deere, D.W., 1989. Rock Quality Designation (RQD) After Twenty Years.

U.S. Army Engineer Waterways Experiment Station Publisher, 34 pp.

Deere, D.U., Miller, R.P., 1966. Engineering classification and index properties for intact

rock. Air Force Weapons Laboratory Technical Report AFWL-TR-65-116, 277 pp.

Demetriades, A., 2014. Basic Considerations: Sampling, the Key for a Successful Applied

Geochemical Survey for Mineral Exploration and Environmental Purposes.

Reference Module in Earth Systems and Environmental Sciences 15: Analytical

Geochemistry/Inorganic INSTR. Analysis, 1–31.

Dershowitz, W.S., Einstein, H.H., 1988. Characterizing rock joint geometry with joint

system models. Rock Mechanics and Rock Engineering 21, 21–51.

Deutsch, C., Journel, A.G., 1998. GSLIB: Geostatistical Software Library and User's

Guide Second Edition. Oxford University Press, New York.

Dewey, J.F., Pitman, W.C., Ryan, W.B., Bonnin, J., 1973. Plate tectonics and the

evolution of the Alpine system. Geological Society of America Bulletin 84, 3137–

3180.

271
Dimitrakopoulos, R., Farrelly, C.T., Godoy, M., 2002. Moving forward from traditional

optimization: grade uncertainty and risk effects in open-pit design. Mining

Technology 111 (A), 82–8.

Dimitrakopoulos, R., Martinez, L., Ramazan, S., 2007. A maximum upside/minimum

downside approach to the traditional optimization of open pit mine design. Journal

of Mining Science 43, 73–82.

Dincer, T., 2001. Application of Pit Optimisation Algorithms beyond Open Pit Limits. 17 th

International Mining Congress and Exhibition of Turkey- IMCET 2001, ISBN 975-

395, 417–4.

Dowd, P., 1994. Risk assessment in reserve estimation and open pit planning.

Transactions of the Institution of Mining and Metallurgy 103, A148–A154.

Dowd, P.A., Onur, A.H., 1992. Optimizing open pit design and scheduling. 23rd APCOM

symposium, 411-422.

Dowd, P., Onur, A., 1993. Open pit optimisation: I. Optimal open pit design. Transactions

of the American Institute of Mining and Metallurgical Engineers, A94–A104.

Efron, B., 1982. The Jackknife, the Bootstrap, and other Resampling Plans. Society for

Industrial and Applied Math, Philadelphia.

Ehlen, J., 2000. Fractal analysis of joint patterns in granite. International Journal of Rock

Mechanics and Mining Sciences 37, 909-922.

Elkington, T., Durham, R., 2011. Integrated open pit pushback selection and production

capacity optimization. Journal of Mining Science 47 (2), 177-190.

Emery, X., 2012. Análisis estadístico de datos. FACULTAD DE CIENCIAS, FISICAS Y

MATEMATICAS, UNIVERSIDAD DE CHILE, 30 pp (In Spanish).

272
Emery, X., Ortiz, J.M., 2005. Histogram and variogram inference in the multi Gaussian

model. Stochastic Environmental Research and Risk Assessment 19 (1), 48–58.

Emery, X., Ortiz, J.M., 2011. Two approaches to direct block-support conditional co-

simulation. Computers & Geosciences 37 (8), 1015–1025.

Englund, E.J., Weber, D.D., Leviant, N., 1992. The effects of sampling design parameters

on block selection. Mathematical Geology 24, 29–343.

Eslamian, S., 2014. Handbook of Engineering Hydrology: Fundamentals and

Applications. CRC Press, 218 pp.

Evertz, C.J.G., Mandelbrot, B.B., 1992. Multifractal measures (appendix B). In: Peitgen,

H. O., Jurgens, H., Saupe, D. (Eds.), Chaos and Fractals. Springer, New York.

Falconer, K., 1991. Fractals. Non-Integral Dimensions and Applications, G. Gherbit (Ed.),

John Wiley, New York, 249. Endeavour 15 (2), 95 pp.

Falconer, K., Hu, J., 2001. Nonlinear Diffusion Equations on Unbounded Fractal Domains.

Journal of Mathematical Analysis and Applications 256 (2), 606–624.

Fisher, R. A., Tippett, L. H. C., 1928. Limiting forms of the frequency distribution of the

largest or smallest member of a sample. In: Mathematical Proceedings of the

Cambridge Philosophical Society 24(2), pp. 180-190.

Fletcher, W.K., 1981. Analytical methods in geochemical exploration. G.J.S. Govett

(Editor), Handbook of Exploration Geochemistry 1, Elsevier, Amsterdam, 25-46.

Frangois-Bongarcon, D., Guidal, D., 1982. Algorithms for parametrizing reserves under

different geometrical constraints. In: T.B. Johnson and R.J. Barnes (Eds.),

Proceedings of 17th APCOM Symposium, Society of Mining and Engineering. AIME,

New York, NY, pp. 297-309.

273
Franke, R., 1982. Scattered data interpolation: tests of some methods. Mathematics of

Computation 38, 181–200.

Gałuszka, A., 2007. A review of geochemical background concepts and an example using

data from Poland. Environmental Geology, 52, 861–870.

Godoy, M.C., Dimitrakopoulos, R., 2004. Managing risk and waste mining in long-term

production scheduling. SME Transactions, 316 pp.

Goncalves, M.A., Mateus, A., Oliveira, V., 2001. Geochemical anomaly separation by

multifractal modeling. Journal of Geochemical Exploration 72, 91–114.

Goodwin, G.C., Seron, M.M., Mayne, D.Q., 2008. Optimization opportunities in mining,

metal and mineral processing. Annual Reviews in Control 32, 17–32.

Goovaerts, P., 1997. Geostatistics for Natural Resources Evaluation. Oxford University

Press, New York, 496 pp.

Grenon, M., Hadjigeorgiou, J., 2010. Integrated structural stability analysis for preliminary

open pit design. International Journal of Rock Mechanics and Mining Sciences 47

(3), 450–460.

Gu, X.W., Wang, Q., Chu, D.Z., Zhang, B., 2010. Dynamic optimization of cutoff grade in

underground metal mining. Journal of Central South University of Technology 17

(3), 492-497.

Halsey, T.C., Jensen, M.H., Kadanoff, L.P., Procaccia, I., Shraiman, B.I., 1986. Fractal

measures and their singularities: the characterization of strange sets. Physical

Review, A 33 (2), 1141–1151.

274
Hamdi, E., Mouza, JD., 2005. A methodology for rock mass characterisation and

classification to improve blast results. International Journal of Rock Mechanics and

Mining Sciences 42, 177–94.

Harati, H., Khakzad, A., Rashidnejad Omran, N., Afzal, P., Hosseini, M., Harati, S., 2013.

Identifying Hydrothermal Alteration: Geochemical Particulars based on

Lithogeochemical Data from the Kahang Cu Porphyry Deposit, Central Iran. Iranian

Journal of Earth Sciences 5, 1-12.

Harrison, J.P., 1999. Selection of the threshold value in RQD assessments. International

Journal of Rock Mechanics and Mining Sciences 36, 673-685.

Hartley, A.J., Rice, C.M., 2005. Controls on supergene enrichment of porphyry copper

deposits in the Central Andes—A review and discussion: Mineralium Deposita 40,

515–525.

Hassanpour, S., Afzal, P., 2013. Application of concentration–number (C–N) multifractal

modeling for geochemical anomaly separation in Haftcheshmeh porphyry system,

NW Iran. Arabian Journal of Geosciences 6 (3), 957–970.

Hawkes, H.E., Webb, J.S., 1962. Geochemistry in Mineral Exploration. Harper and Row,

New York.

He, Y., Zhu, K., Gao, S., Liu, T., Li, Y., 2009. Theory and method of genetic-neural

optimizing cut-off grade and grade of crude ore. Expert Systems with Applications

36 (4), 7617–7623.

Heidari, S.M., Ghaderi, M., Afzal, P., 2013. Delineating mineralized phases based on

lithogeochemical data using multifractal model in Touzlar epithermal Au–Ag (Cu)

deposit, NW Iran. Applied Geochemistry 31, 119–132.

275
Hejazi, M., 2005. Geometry in nature and Persian architecture. Building and Environment

40, 1413–1427.

Hitzman, M.W., Oreskes, N., Einaudi, M.T., 1992. Geological characteristics and tectonic

setting of Proterozoic iron oxide (Cu-U-Au-REE) deposits. Precambrian Research

58, 241-287.

Hochbaum, D.S., Chen, A., 2000. Performance analysis and best implementations of old

and new algorithms for the open-pit mining problem. Operations Research 48, 894–

913.

Homayoon, SR., Keshavarzi, A., Gazni, R., 2010. JSEA Application of artificial neural

network, kriging, and inverse distance weighting models for estimation of scour

depth around bridge pier with bed sill. Journal of Software Engineering and

Applications 3, 944–964.

Howarth, R.J., Thompson, M., 1976. Duplicate analysis in geochemical practice, Part II.

Analyst 101, 699-709.

Hustrulid, W., Kuchta, M., 2006. Open Pit Mine Planning and Design. Taylor & Francis.

Isaaks, E., Srivastava. R., 1989. An Introduction to Applied Geostatistics. Oxford

University Press, New York.

Jinga, L., Hudson, J.A., 2002. Numerical methods in rock mechanics. International

Journal of Rock Mechanics and Mining Sciences 39, 409–427.

Johnson, T B., 1968. Optimum open pit mine production scheduling in mining

engineering, University of California: Berkley, 120 pp.

276
Juan, P., Mateu, J., Jordan, M.M., Mataix-Solera, J., Meléndez-Pastor, I., Navarro-

Pedreño, J., 2011. Geostatistical methods to identify and map spatial variations of

soil salinity. Journal of Geochemical Exploration 108, 62–72.

Julia, G., 1959. Éléments d'algèbre. Gauthier-Villars, Paris, 1-207 (In French).

Journel, A., 1993. Geostatistics: roadblocks and challenges, In A. Soares, (Ed.),

Geostatistics-Troia, pp. 213-224.

Journel, A.G., Huijbregts, Ch. J., 1978. Mining Geostatistics, Virginia, Academic Press,

600 pp.

Kalenchuk, K.S., Diederichs, M.S., McKinnon, S., 2006. Characterizing block geometry

in jointed rock masses. International Journal of Rock Mechanics and Mining

Sciences 43, 1212–25.

Kazi, A., Sen, Z., 1985. Volumetric RQD: an index of rock quality, in Fundamentals of

rock joints. In Stephansson (Ed.), Centek Publisher, Sweden. pp. 95-101.

Khayrollahi, H., 2003. The Investigation of Mass Transfer and Element Mobility within the

Sar-Cheshmeh Copper Mine Alteration Zones. (MSc Thesis) Amirkabir University of

Technology, 145 pp.

King, B., 2011. Optimal mining practice in strategic planning. Journal of Mining Science

47 (2), 247-253.

Kirkham, R.V., Dunne, K.P., 2000. World distribution of porphyry, porphyry-associated

skarn, and bulk-tonnage epithermal deposits and occurrences. Geological Survey

of Canada 3792, 1–26.

Koenigsberg, E., 1982. The optimum contours of an open pit mine: an application of

dynamic programming. In: T.B. Johnson and R.J. Barnes (Eds.), Proceedings of 17th

277
 APCOM Symposium, Society of Mining and Engineering. AIME, New York, NY, pp.

274-287.

Krautkraemer, J.A., 1988. The cut-off grade and the theory of extraction. Canadian

Journal of Economics 21 (1), 146–160.

Krige, D.G., 1966. A study of gold and uranium distribution patterns in the Klerksdorp

goldfield. Geoexploration 4, 43–53.

Lam, N.S., 1983. Spatial interpolation methods: A review. American Cartographer 10,

129–149.

Lane, K.F., 1964. Choosing the optimum cut-off grade. Quarterly of the Colorado School

of Mines, Golden, USA, pp. 811–829.

Lauwerier, H., 1991. Fractals: Images of Chaos. Princeton University Press, New Jersey.

240 pp.

Laznicka, P., 2005. Giant Metallic Deposits Future Sources of Industrial Metals. Springer-

Verlag. 732 pp.

Lerch, H., Grossmann, I.F., 1965. Optimum design of open-pit mines. Canadian Institute

of Mining and Metallurgy 58, 47–54.

Li, C., Ma, T., Shi, J., 2003. Application of a fractal method relating concentrations and

distances for separation of geochemical anomalies from background. Journal of

Geochemical Exploration 77, 167–175.

Lima, A., Albanese, S., Cicchella, D., 2005. Geochemical baselines for the radioelements

K, U, and Th in the Campania region, Italy: a comparison of stream–sediment

geochemistry and gamma-ray surveys. Journal of Geochemical Exploration 20,

611–625.

278
Lima, A., De Vivo, B., Cicchella, D., Cortini, M., Albanese, S., 2003. Multifractal IDW

interpolation and fractal filtering method in environmental studies: an application on

regional stream sediments of (Italy), Campania region. Applied Geochemistry 18,

53–65.

Lima, A., Plant, J.A., Vivo, B.D., Tarvainen, T., Albanese, S., Cicchella, D., 2008.

Interpolation methods for geochemical maps: a comparative study using arsenic

data from European stream waters. Geochemistry: Exploration, Environment,

Analysis 8, 41–48.

Limpert, E., Stahel,W.A., Abbt,M., 2001. Lognormal distributions across the sciences:

keys and clues. Bioscience 51 (5), 341–352.

Lina, J.S., Kub, C.Y., 2006. Two-scale modeling of jointed rock masses. International

Journal of Rock Mechanics and Mining Sciences 43, 426-436.

Little, M., 2006. The benefit to open pit rock slope design of geotechnical databases. In:

Proceedings of international symposium on stability of rock slopes in open pit mining

and civil engineering. SAIMM, pp. 97–116.

London Metal Exchange, 2015a. LME Copper. https://www.lme.com/metals/non-

ferrous/copper/.

London Metal Exchange, 2015b. LME Molybdenum. https://www.lme.com/metals/minor-

metals/molybdenum/.

Lowell, J.D., 1968. Geology of the Kalamazoo orebody, San Manuel district, Arizona.

Economic Geology 63, 645—654.

Lowell, J.D., Guilbert, J.M., 1970. Lateral and vertical alteration-mineralization zoning in

porphyry ore deposits. Economic Geology 65, 373–408.

279
Luz, F., Mateus, A., Matos, J.X., Gonçalves,M.A., 2014. Cu- and Zn-Soil Anomalies in the

NE Border of the South Portuguese Zone (Iberian Variscides, Portugal) Identified

by Multifractal and Geostatistical Analyses. Natural Resources Research 23, 195–

215.

Madani Esfahani, N., Asghari, O., 2013. Fault detection in 3D by sequential Gaussian

simulation of Rock Quality Designation (RQD), Arabian Journal of Geosciences 6,

3737–3747.

Makkonen, H., Mäkinen, J., Kontoniemi, O., 2008. Geochemical discrimination between

barren and mineralized intrusions in the Svecofennian (1.9 Ga) Kotalahti Nickel Belt,

Finland. Ore Geology Reviews 33, 101–114.

Mandelbrot, B.B., 1983. The Fractal Geometry of Nature. W.H. Freeman, San Francisco,

CA. Updated and Augmented Edition.

Marcotte, D., Caron, J., 2013. Ultimate open pit stochastic optimization. Computers &

Geosciences 51, 238–246.

Mart, W.S., Markey, G., 2013. Intelligent Mining Software “Solutions” IMS - Lerch-

Grossman Pit Optimization. MineMap Pty Ltd.

McInnes, B.I.A., Evans, N.J., Fu, F.Q., Garwin, S., Belousova, E., Griffin, W.L., Bertens,

A., Sukama, D., Permanadewi, S., Andrew, R.L., Deckart, K., 2005. Thermal history

analysis of selected Chilean, Indonesian, and Iranian porphyry Cu–Mo–Au deposits.

In: Porter, T.M. (Ed.), Super Porphyry Copper and Gold Deposits: A Global

Perspective. PGC publishing, Adelaide. pp 1–16.

280
Meagher, C., Avis, D., Dimitrakopoulos, R., 2010. The Direct Cut Polytope, Integrality and

Triangular Elimination: Theoretical developments for mining optimization problems,

COSMO Research Report no. 5, 64-88.

Melfos, V., Vavelidis, M., Christo des, G., Seidel, E., 2002. Origin and evolution of the

Tertiary Maronia porphyry copper-molybdenum deposit, Thrace, Greece.

Mineralium Deposita 37, 648—668.

Meyer, C., Hemley, J. J., 1967. Wall rock alteration in barnes. H. L. (Ed.) Geochemistry

of hydrothermal ore deposits, Holt, Reinhart and winston, New York. pp. 166-235.

Meyer, T., Einstein, H.H., 2002. Geologic stochastic modelling and connectivity

assessment of fracture systems in the Boston area. Rock Mechanics and Rock

Engineering 35, 23–44.

Mihalasky, M.J., Bookstrom, A.A., Frost, T.P., Ludington, S., with contributions from

Logan, J.M., Panteleyev, A., Abbot, G., 2011, revised 2013. Porphyry copper

assessment of British Columbia and Yukon Territory, Canada: U.S. Geological

Survey Scientific Investigations Report 2010-5090-C, v. 1.1, 128 pp.

Milne, D., 2007. Rock mass classification challenges. Rock Mechanics: Meeting Society's

Challenges and Demands, In: Proceedings of the 1st Canada-US Rock Mechanics

Symposium, Vancouver, Canada, pp. 191-197.

Mirnejad, H., Mathur, R., Einali, M., Dendas, M., Alirezaei, S., 2010. A comparative

copper isotope study of porphyry copper deposits in Iran. Geochemistry:

Exploration, Environment, Analysis 10, 413–418.

281
Monecke, T., Gemmell, J.B., Monecke, J., 2001. Fractal distributions of veins in drill core

from the Hellyer VHMS deposit, Australia: constraints on the origin and evolution of

the mineralising system. Mineralium Deposita 36 (5), 406-415.

Monecke, T., Monecke, J., Herzig, P.M., Gemmell, J.B., Monch, W., 2005. Truncated

fractal frequency distribution of element abundance data: a dynamic model for the

metasomatic enrichment of base and precious metals. Earth and Planetary Science

Letters 232, 363–378.

Montiel, L., Dimitrakopoulos, R., 2013. Stochastic mine production scheduling with

multiple processes: Application at Escondida Norte, Chile. Journal of Mining

Science 49 (4), 583-597.

Nash, J.T., 1976. Fluid inclusion petrology-data from porphyry copper deposits and

applications to exploration. Geological Survey Professional Paper 907-D, 1—16.

Neimark, J.I., 2003. Mathematical Models in Natural Science and Engineering: An

Example-Based Approach. Springer Science & Business Media Publisher, Berlin-

Germany.

Nematzadeh, M.R., 2015. Speech of minister of Industries, Mines and Commerce of Iran

at the Parliament. www.dolat.ir/NSite/FullStory/Cabine/?Serv=0&Id=208570

(Government Information Centre: In Persian).

NPV – Scheduler. 2001. Tutorial of Earthworks NPV-Scheduler software version 3.4.

Copyright @ 2001 Mineral Industries Computing Ltd.

Olea, R.A., 2007. Declustering of clustered preferential sampling for global histogram and

semi variogram inference. Mathematical Geology 39(5), 453–467.

282
Oliveira Neto, M., 2006. Pythagoras’ celestial spheres in the context of a simple model

for quantization of planetary orbits. Chaos, Solitons and Fractals 30, 399–406.

Ortega, O.J., Marrett, R., Laubach, S.E., 2006. A scale-independent approach to fracture

intensity and average fracture spacing. American Association of Petroleum

Geologists Bulletin 90 (2), 193–208.

Osanloo, M., Ataei, M., 2003. Using equivalent grade factors to find the optimum cut-off

grades of multiple metal deposits. Minerals Engineering 16, 771–776.

Osanloo, M., Gholamnejad, J., Karimi, B., 2008a. Long-term open pit mine production

planning: a review of models and algorithms. International Journal of Mining,

Reclamation and Environment 22, 3–35.

Osanloo, M., Rashidinejad, F., Rezai, B., 2008b. Incorporating environmental issues into

optimum cut-off grades modeling at porphyry copper deposits. Resources Policy 33,

222–229.

Palmstrom, A. 1985. Application of the volumetric joint count as a measure of rock mass

jointing, in Fundamentals of rock joints. In Stephansson (Ed.), Centek Publisher,

Sweden. pp. 103-109.

Palmstrom, A., 2005. Measurements of and Correlations between Block Size and Rock

Quality Designation (RQD). Tunnels and Underground Space Technology 20, 362-

377.

Panahi, A., Cheng, Q., Bonham-Carter, G.F., 2004. Modelling lake sediment geochemical

distribution using principal component, indicator kriging and multifractal power

spectrum analysis: a case study from Gowanda, Ontario. Exploration, Environment,

Analysis 4, 59–70.

283
Parhizkar, A., Ataei, M., Moarefvand, P., Rasouli, V., 2011. Grade uncertainty and its

impact on ore grade reconciliation between the resource model and the mine.

Archives of Mining Sciences 56119–134.

Parhizkar, A., Ataei, M., Moarefvand, P., Rasouli, V., 2012. A probabilistic model to

improve reconciliation of estimated and actual grade in open-pit mining. Arabian

Journal of Geosciences 5 (6), 1279-1288.

Picard, J. C., 1976. Maximal closure of a graph and applications to combinatorial

problems. Management Science 22, 1268–1272.

Pirajno, F., 2009. Hydrothermal processes and mineral system: Vol. 1, Springer, 1250 pp.

Priest, S.D., Hudson, J.A., 1976. Discontinuity spacing in rock. International Journal of

Rock Mechanics and Mining Sciences- Geomechanics Abstracts 13, 135-148.

Plant, J., Smith, D., Smith, B., Williams, L., 2001. Environmental geochemistry at the

global scale. Applied Geochemistry 16, 1291–1308.

Popov, K.P., Ruskov, K.I., Georgiev, G.I., 2003. 3D Geostatistical model of the ore body

in Elatsite porphyry copper deposit, Panagyurishte ore region. 50 years University

of Mining and Geology “St. Ivan Rilski” Annual 46 (I), Geology and Geophysics,

Sofia, 113-118.

Rafiee, A., Vinches, M., 2008. Application of geostatistical characteristics of rock mass

fracture systems in 3D model generation. International Journal of Rock Mechanics

and Mining Sciences 45, 644–52.

Ramazan, S., 2007. The new Fundamental Tree Algorithm for production scheduling of

open pit mines, European Journal of Operational Research 177, 1153–1166.

284
Razique, A., Lo Grasso, G., Livesey, T., 2007. Porphyry Copper–Gold Deposits at Reko

Diq Complex, Chagai Hills Pakistan. In: Proceedings of Ninth Biennial SGA Meeting,

Dublin, pp. 1–7.

Reimann, C., Filzmoser, P., 2000. Normal and lognormal data distribution in

geochemistry: death of a myth. Consequences for the statistical treatment of

geochemical and environmental data. Environmental Geology 39, 1001–1014.

Richards, J.P., Spell, T., Rameh, E., Razique, A., Fletcher, T., 2012. High Sr/Y magmas

reflect arc maturity, high magmatic water content, and porphyry Cu ± Mo ± Au

potential: examples from the Tethyan arcs of Central and Eastern Iran and Western

Pakistan. Economic Geology 107, 295–332.

Richmond, A., 2002. Two-point declustering for weighting data pairs in experimental

variogram calculations. Computers & Geosciences 28,231–241.

Robb, L., 2005. Introduction to Ore-forming processes, Blackwell Science Ltd, 325 pp.

RockWorks 15 by RockWare®, Inc. 2010. Tutorial of RockWorks 15TM, Third edition.

Roedder, E., 1971. Fluid inclusion studies on the porphyry-type ore deposits at Bingham,

Utah, Butte, Montana, and Climax, Colorado. Economic Geology 66, 98—120.

Rouleau, A., Gale, J.E., 1985. Statistical characterisation of fracture system in the Stripa

Granite, Sweden. International Journal of Rock Mechanics and Mining Sciences-

Geomechanics Abstracts 22(6), 353-367.

Sadeghi, B., Madani, N., Carranza, E.J., 2015. Combination of geostatistical simulation

and fractal modeling for mineral resource classification. Journal of Geochemical

Exploration 149, 59–73.

285
Sadeghi, B., Moarefvand, P., Afzal, P, Yasrebi, A.B., Daneshvar Saein, L., 2012.

Application of fractal models to outline mineralized zones in the Zaghia iron ore

deposit, Central Iran. Journal of Geochemical Exploration 122, 9–19.

Sameni Keivani, F., Khalili Sourkouhi, Z., 2014. A Survey on Supply Side of Mines: a

Case Study of Iran. European Online Journal of Natural and Social Sciences 3, 589-

596.

Schroder, J.W., 1944. Essaisur la structure de l'Iran. Ecologae Geologicae Helvetiae 37,

37-81 (In French).

Schwab, R., Ruff, V., 2013. On the nature of the yield point phenomenon. Acta Materialia

61, 1798–1808.

Schwartz, G.M., 1947. Hydrothermal alteration in the “porphyry copper” deposits.

Economic Geology 42, 319–352.

Şen, Z., Eissa, E.A., 1991. Volumetric rock quality designation. Journal of Geotechnical

Engineering (ASCE) 117, 1331–1348.

Şen, Z., Eissa, E.A., 1992. Rock quality charts for log-normally distributed block sizes.

International Journal of Rock Mechanics and Mining Sciences- Geomechanics

Abstracts 29 (1), 1–12.

Seymour, F., 1995. Pit limit parameterisation for modified 3D Lerch-Grossmann

Algorithm. SME Transactions 298, 1860-1864.

Shafiei, B., Haschke, M., Shahabpour, J., 2009. Recycling of orogenic arc crust triggers

porphyry Cu mineralization in Kerman Cenozoic arc rocks, southeastern Iran.

Mineralium Deposita 44, 265–283.

286
Shahabpour, J., 1982. Aspects of Alteration and Mineralization at the Sar-Cheshmeh

Copper Molybdenum Deposit, Kerman, Iran. (PhD Thesis) Leeds University, 342

pp.

Shahabpour, J., 1994. Post-mineral breccia dyke from the Sar-Cheshmeh porphyry

copper deposit, Kerman, Iran. Exploration and Mining Geology 3, 39-43.

Shahbeik, Sh., Afzal, P., Moarefvand, P., Qumarsy, M., 2014. Comparison between

ordinary kriging (OK) and inverse distance weighted (IDW) based on estimation

error. Case study: Dardevey iron ore deposit, NE Iran. Arabian Journal of

Geosciences 7 (9), 3693–3704.

Shen, W., Zhao, P., 2002. Theoretical study of statistical fractal model with applications

to mineral resource prediction. Computers & Geosciences 28, 369–376.

Shepard, D., 1968. A two-dimensional interpolation function for irregularly-spaced data.

In: Proceedings of the 1968 ACM National Conference. pp. 517–524.

Sillitoe, R.H., 1997. Characteristics and controls of the largest porphyry copper–gold and

epithermal gold deposits in the circum-Pacific region. Australian Journal of Earth

Science 44, 373–388.

Sillitoe, R.H., 2005. Supergene oxidized and enriched porphyry copper and related

deposits: Society of Economic Geologists, Economic Geology 100th Anniversary

Volume, 723–768.

Sim, B.L., Agterberg, F.P., Beaudry, C., 1999. Determining the cut-off between

background and relative base metal contamination levels using multifractal

methods. Computers & Geosciences 25, 1023–1041.

287
Singh, V.K., Baliga, B.D., 1994. Slope design of an open pit copper mine. International

Journal of Rock Mechanics and Mineral Science- Geomechanics Abstracts 31, 55–

69.

Slob, S., 2010. Automated Rock Mass Characterisation Using 3-D Terrestrial Laser

Scanning. ITC Publisher, 38 pp.

Soleymani Shishvan, M., Sattarvand, J., 2015. Long term production planning of open pit

mines by ant colony optimization, European Journal of Operational Research 240,

825–836.

Soltani, F., Afzal, P., Asghari, O., 2014. Delineation of alteration zones based on

Sequential Gaussian Simulation and concentration–volume fractal modeling in the

hypogene zone of Sungun copper deposit, NW Iran. Journal of Geochemical

Exploration 140, 64–76.

Soltani Mohammadi, S., Hezarkhan, A., Tercan, E., 2012. Optimally Locating Additional

Drill Holes in Three Dimensions Using Grade and Simulated Annealing. Journal of

Geological Society of India 80, 700-706.

Spalla, M.I., Morotta, A.M., Gosso, G., 2010. Advances in interpretation of geological

processes: refinement of multi-scale data and integration in numerical modelling.

Geological Society, London, 240 pp.

Stermole, F., Stermole, J., 2012. Economic Evaluation and Investment Decisions

Methods Textbook; 13th Ed, Graphic Designer, USA, 159 pp.

Sun, T., Liu, L., 2014. Delineating the complexity of Cu-Mo mineralization in a porphyry

intrusion by computational and fractal modeling: A case study of the Chehugou

288
 deposit in the Chifeng district, Inner Mongolia, China. Journal of Geochemical

Exploration 144, 128–143.

Sweetkind, D.S., Drake, R.M. II, 2007, Geologic characterization of young alluvial basin-

fill deposits from drill-hole data in Yucca Flat, Nye County, Nevada: U.S. Geological

Survey Scientific Investigations Report 2007–5062, 17 pp.

Tabatabaei, S. H., Asadi Haroni, H., 2006. Geochemical characteristics of Gor Gor Cu-

Mo porphyry system, 25th Iranian symposium on geosciences, Geological survey of

Iran, 60 pp.

Tahmasebi, P., Hezarkhani, A., 2010. Application of Adaptive Neuro-Fuzzy Inference

System for Grade Estimation; Case Study, Sarcheshmeh Porphyry Copper Deposit,

Kerman, Iran. Australian Journal of Basic and Applied Sciences 4, 408-420.

Takayasu, H., 1990. Fractals in the Physical Sciences. Manchester University Press,

Manchester and New York. 176 pp.

Thompson, M., Howarth, S.R., 1978. A new approach to the estimation of analytical

precision. Journal of Geochemical Exploration 9, 23-30.

Todorov, J., Popov, K., Shanov, S., Boykova, A. 2002. Geological Conditions for a Correct

Geostatistical Evaluation: Example from the Elatsite Copper Deposit in Bulgaria. In:

Geostatistics Rio 2000, Kluwer Academic Publishers, 177-189.

Trading Economics, 2015a. Iran Interest Rate 1973-2015 Data Chart.

http://www.tradingeconomics.com/iran/interest-rate.

Trading Economics, 2015b. United Kingdom Interest Rate 1971-2015 Data Chart.

http://www.tradingeconomics.com/united-kingdom/interest-rate.

Tukey, J.W., 1977. Exploratory data analysis. Addison-Wesley, Reading, USA.

289
Turcotte, D.L., 1986. A fractal approach to the relationship between ore grade and

tonnage. Economic Geology 18, 1525–1532.

Turcotte, D.L., 1997. Fractals and Chaos in Geology and Geophysics. Cambridge

University Press, Cambridge.

Ulrich, T., Gunther, D., Heinrich, C.A., 2001. The evolution of a porphyry Cu—Au deposit,

based on La-ICP-MS analysis of fluid inclusions, Bajo de la Alumbrera, Argentina.

Economic Geology 96, 1743—1774.

Verly, G., 1984. Estimation of Spatial Point and Block Distributions: The Multi-Gaussian

Model, PhD thesis, Stanford University, Stanford, CA.

Villaescusa, E., Brown, E.T., 1990. Characterising joint spatial correlation using

geostatistical methods, in Rock Joints. In Barton & Stephansson (Eds.), Balkema

publisher, Rotterdam. pp. 115-122.

Wang, Q., Deng, J., Liu, H., Wang, Y., Sun, X., Wan, L., 2011. Fractal models for

estimating local reserves with different mineralization qualities and spatial

variations. Journal of Geochemical Exploration 108, 196–208.

Wang, Q.F., Deng, J., Liu, H., Yang, L.Q., Wan, L., Zhang, R.Z., 2010a. Fractal models

for ore reserve estimation. Ore Geology Reviews 37, 2–14.

Wang, Q., Deng, J., Zhao, J., Li, N., Wan, L., 2012. The fractal relationship between

orebody tonnage and thickness. Journal of Geochemical Exploration 122, 4–8.

Wang, Q., Deng, J., Zhao, J., Liu, H., Wan, L., Yang, L., 2010b. Tonnage-cut-off model

and average grade-cut-off model for a single ore deposit. Ore Geology Reviews 38

(1–2), 113–120.

290
Wang, G., Pang, Z., Boisvert, J.B., Hao, Y., Cao, Y., Qu, J., 2013. Quantitative

assessment of mineral resources by combining geostatistics and fractal methods in

the Tongshan porphyry Cu deposit (China). Journal of Geochemical Exploration

134, 85–98.

Weber, D.D., Englund, E.J., 1992. Evaluation and comparison of spatial interpolators.

Mathematical Geosciences 24, 381–391.

Weber, D.D., Englund, E.J., 1994. Evaluation and comparison of spatial interpolators, II.

Mathematical Geology 26, 589–603.

Whittle, J., 1988. Beyond Optimization in Open Pit Design, Proc. 1st Canadian Conference

on Computer Applications in the Mineral Industry, Quebec City, pp. 331-337.

Whittle, J., 1989. The Facts and fallacies of open pit optimization. Whittle Programming

Pty Ltd.

Whittle Programming, 1998a. Four-X, Reference Manual, Whittle Programming,

Melbourne, 385 pp.

Whittle, J., 1998b. Beyond optimization in open pit design. In: Proceedings of the

Canadian Conference on Computer Applications in the Mineral Industries, pp. 331-

337.

Whittle, J., 1990. Surface Mining Society for Mining Metallurgy and Exploration (Open Pit

Optimization), Colorado, pp. 470– 475.

Whittle, J., 1999. A decade of open pit mine planning and optimisation - the craft of turning

algorithms into packages. In: Proceedings of the 28th International Symposium on

Application of Computers and Operations Research in the Mineral Industry, pp. 15-

24.

291
Wilson, A.J., Cooke David, R., Harper, B.J., Deyell, C.L., 2007. Sulfur isotopic zonation

in the Cadia district, southeastern Australia: exploration significance and

implications for the genesis of alkalic porphyry gold-copper deposits. Mineralium

Deposita 42, 465—487.

Wyllie, D.C., Mah, C.W., 2004. Rock slope engineering. London, Taylor & Francis.

Xiao, F., Chen, j., Agterberg, F., Wang, C., 2014. Element behavior analysis and its

implications for geochemical anomaly identification: A case study for porphyry Cu–

Mo deposits in Eastern Tianshan, China. Journal of Geochemical Exploration 145,

1–11.

Xu, Y., Cheng, Q., 2001. A fractal filtering technique for processing regional geochemical

maps for mineral exploration. Geochemistry: Exploration, Environment, Analysis 1,

147–156.

Yamatomi, J., Mogi, G., Akaike, A., Yamaguchi, U., 1995. Selective extraction dynamic

cone algorithm for three dimensional open pit designs. 25th APCOM Conference,

pp. 267-274.

Yang, L., Wang, Q., Liu, X., 2015. Correlation between mineralization intensity and fluid–

rock reaction in the Xinli gold deposit, Jiaodong Peninsula, China: Constraints from

petrographic and statistical approaches. Ore Geology Reviews 71, 29–39.

Yasrebi, A.B., Afzal, P., Wetherelt, A., Foster, P., Esfahanipour, R., 2013a. Correlation

between geology and concentration-volume fractal models: significance for Cu and

Mo mineralized zones separation in the Kahang porphyry deposit (Central Iran).

Geologica Carpathica 64, 153–163.

292
Yasrebi, A.B., Afzal, P., Wetherelt, A., Foster, P., Esfahanipour, R., Parvaz, D., 2012. 3D

Lithology, Alteration and Zonation Modelling of the Eastern Part of Kahang Cu-Mo-

Au Porphyry Deposit in Central Iran. 12th International Multidisciplinary Scientific

GeoConference SGEM, Varna, pp. 417-424.

Yasrebi, J., Saffari, M., Fathi, H., Karimian, N., Moazallahi, M., Gazni, R., 2009.

Evaluation and comparison of ordinary kriging and inverse distance weighting

methods for prediction of spatial variability of some chemical parameters. Research

Journal of Biological Sciences 4, 93–102.

Yasrebi, A.B., Wetherelt, A., Foster, P., Afzal, P., 2011. Determination and analysis of

final pit limit of Esfordi phosphate open pit mine. 22th World Mining Congress,

Turkey, pp. 513–522.

Yasrebi, A.B., Wetherelt, A., Foster, P., Afzal, P., Coggan, J., Kaveh Ahangaran, D.,

2013b. Application of RQD-Number and RQD-Volume Multifractal Modeling to

delineate rock mass characterisation in Kahang Cu-Mo Porphyry Deposit, Central

Iran. Archives of Mining Sciences 58 (4), 1023–1035.

Yasrebi, A.B., Wetherelt, A., Foster, P., Coggan, J., Afzal, P., Agterberg, F., Kaveh

Ahangaran, D., 2014. Application of a density–volume fractal model for rock

characterisation of the Kahang porphyry deposit. International Journal of Rock

Mechanics and Mining Sciences 66, 188-193.

Young, K.C., 1978. Ultimate pit design methodologies using computer models-the state

of the art. AIME Transactions (264), 1454-1459.

Yousefi, M., Kamkar-Rouhani, A., Carranza, E.J.M., 2012. Geochemical mineralization

probability index (GMPI): A new approach to generate enhanced stream sediment

293
 geochemical evidential map for increasing probability of success in mineral potential

mapping. Journal of Geochemical Exploration 115, 24–35.

Yousefi, M., Kamkar-Rouhani, A., Carranza, E.J.M., 2014. Application of staged factor

analysis and logistic function to create a fuzzy stream sediment geochemical

evidence layer for mineral prospectivity. Geochem. Geochemistry: Exploration,

Environment, Analysis 14, 45–58.

Zarasvandi, A., Liaghat, S., Zentilli, M., 2005. Geology of the Darreh–Zerreshk and Ali-

Abad porphyry copper deposit, central Iran. International Geology Review 47, 620–

646.

Zeng, Q., Liu, J., Zhang, Z., Chen, W., Qin, F., Zhang, R., Yu, W., Zhang, X., Zhai, M.,

2009. Mineralizing type, geological characteristics and geodynamic background of

molybdenum deposit in Xilamulun molybdenum polymetal metallogenic belt on

northern margin of North China Craton. Acta Petrolei Sinica 25 (5), 1225–1238 (in

Chinese with English abstract).

Zhang, C., Jordan, G., Higgins, A., 2007. Using neighbourhood statistics and GIS to

quantify and visualize spatial variation in geochemical variables: an example using

Ni concentrations in the topsoils of Northern Ireland. Geoderma 137, 466–476.

Zhang, L., Einstein, H.H., 2000. Estimating the intensity of rock discontinuities.

International Journal of Rock Mechanics and Mining Sciences 37, 819–37.

Zhang, W., Wang, Q., Chen, J.P., Tan, C., Yuan, X.Q., Zhou, F.J., 2012. Determination

of the optimal threshold and length measurements for RQD calculations,

International Journal of Rock Mechanics and Mining Sciences 51, 1–12.

294
Zhao, Y., Kim, Y.C., 1992. A New Ultimate Pit Limit Design Algorithm. 23rd APCOM, 423-

434.

Zhmud, L., 1989. Pythagoras as a mathematician. Historia Mathematica 16, 249-268.

Zhu, Z., Jia, B., Zhou, Z., 2003. The lower bound of the Hausdorff measure of Koch Curve,

Acta Mathematica Sinica, English series 19, 106–112.

Zimmerman, D., Pavlik, C., Ruggles, A., Armstrong, M.P., 1999. An Experimental

comparison of ordinary and universal kriging and inverse distance weighting.

Mathematical Geology 31, 375–390.

Zuo, R., 2014. Identification of weak geochemical anomalies using robust neighborhood

statistics coupled with GIS in covered areas. Journal of Geochemical Exploration

136, 93–101.

Zuo, R., Carranza, E.J.M., Cheng, Q., 2012. Fractal/multifractal modelling of geochemical

exploration data. Journal of Geochemical Exploration 122, 1–3.

Zuo, R., Cheng, Q., Agterberg, F.P., Xia, Q., 2009b. Application of singularity mapping

technique to identify local anomalies using stream sediment geochemical data, a

case study from Gangdese, Tibet, western China. Journal of Geochemical

Exploration 101, 225–235.

Zuo, R., Cheng, Q., Xia, Q., 2009a. Application of fractal models to characterization of

vertical distribution of geochemical element concentration. Journal of Geochemical

Exploration 102, 1, 37—43.

Zuo, R., Wang, J., 2015. Fractal/multifractal modeling of geochemical data: A review.

Journal of Geochemical Exploration (In Press, Corrected Proof).

295
Zuo, R., Wang, J., Chen, G., Yang, M., 2015. Identification of weak anomalies: A

multifractal perspective. Journal of Geochemical Exploration 148, 12–24.

296
Appendices

297
Appendix. A: Certificate of Data Sources

298
Fig. A. 1. Letter issued by the Kahang deposit project manager authorising the use of data and
choice of input parameters to the pit optimisation study

299
 Appendix. B: Cu Re-Analysed or Duplicated Samples in the Kahang
Deposit
Table. B. 1. 399 randomised samples for Cu selected and analysed for assay quality assurance
and quality control

300
301
302
303
304
305
306
307
 Appendix. C: Geochemical Data Variances via F-Distribution

Table. C. 1. Fisher distribution F(n1,n2) with n1 and n2 degrees of freedom, α = 0.025 and 97. 5%
of confidence level (Emery, 2012)

308
 Appendix. D: Distribution of Student (T) with n Degrees of Freedom

Table. D. 1. Critical values for student’s T distributions (column headings denote probabilities’ α
above tabulated values: Emery, 2012)

309
Appendix. E: Density and RQD Plans in Different Levels

Fig. E. 1. Density plan view for layer # 32 (Z =1900 m)

310
Fig. E. 2. RQD plan view for layer # 32 (Z =1900 m)

311
Fig. E. 3. Density plan view for layer # 37 (Z =1950 m)

312
Fig. E. 4. RQD plan view for layer # 37 (Z =1950 m)

313
Fig. E. 5. Density plan view for layer # 42 (Z =2000 m)

314
Fig. E. 6. RQD plan view for layer # 42 (Z = 2000m)

315
 Appendix. F: Economic Models in the Kahang Deposit
Table. F. 1. The individual economic properties at all elevations (generated by NPV Scheduler
software)

Rock Revenue Processing Cost Mining Cost ORE Cu Mo Cu R Mo R

tonnes $ $ $ tonnes tonnes g tonnes g
Elevation 2360.00 9,023,760 0 0 36,095,040 0 0 0 0 0
Elevation 2350.00 9,023,760 0 0 36,095,040 0 0 0 0 0
Elevation 2340.00 9,032,698 352,689 235,197 36,543,325 56,520 57.924 2,131,657 44.949 1,654,166
Elevation 2330.00 9,181,637 3,933,008 2,573,962 38,051,538 616,757 694.2483 9,068,299 538.7367 7,037,000
Elevation 2320.00 9,330,494 22,333,601 11,305,510 41,238,636 2,673,549 3,953 48,381,836 3,067 37,544,305
Elevation 2310.00 9,331,390 9,282,607 6,258,133 40,457,095 1,502,934 1,585 37,806,669 1,230 29,337,975
Elevation 2300.00 9,331,123 8,806,450 5,894,721 39,925,122 1,414,317 1,537 25,560,388 1,193 19,834,861
Elevation 2290.00 9,330,651 13,233,657 7,228,669 40,267,129 1,717,344 2,319 35,601,503 1,800 27,626,766
Elevation 2280.00 9,421,616 15,084,434 6,338,199 40,311,682 1,481,192 2,664 34,448,430 2,067 26,731,982
Elevation 2270.00 9,448,712 17,154,578 8,066,531 41,150,138 1,899,480 3,022 41,388,101 2,345 32,117,166
Elevation 2260.00 9,481,392 25,883,725 8,681,095 41,081,036 1,991,168 4,617 44,945,946 3,583 34,878,054
Elevation 2250.00 9,482,797 26,068,428 11,226,408 41,806,517 2,629,322 4,556 73,927,805 3,536 57,367,977
Elevation 2240.00 9,482,554 22,353,675 9,052,246 41,639,910 2,109,890 3,945 51,849,770 3,061 40,235,421
Elevation 2230.00 9,489,563 25,217,308 11,182,161 41,957,644 2,622,562 4,465 53,952,212 3,465 41,866,917
Elevation 2220.00 9,490,387 15,271,558 6,967,022 41,207,374 1,636,674 2,718 28,500,705 2,109 22,116,547
Elevation 2210.00 9,490,979 16,565,029 7,602,964 41,138,049 1,788,267 2,893 47,790,545 2,245 37,085,463
Elevation 2200.00 9,488,928 16,645,859 8,915,088 41,210,997 2,117,070 2,868 59,672,108 2,226 46,305,555
Elevation 2190.00 9,489,166 19,906,548 8,808,052 41,476,358 2,066,827 3,474 57,998,104 2,696 45,006,528
Elevation 2180.00 9,574,547 18,599,586 8,613,910 41,454,441 2,030,506 3,126 90,824,624 2,426 70,479,908
Elevation 2170.00 9,601,150 18,312,328 8,689,065 42,000,570 2,047,771 3,206 50,338,290 2,488 39,062,513
Elevation 2160.00 9,602,128 21,369,488 10,784,993 42,268,380 2,551,000 3,745 57,452,098 2,906 44,582,828
Elevation 2150.00 9,621,101 25,711,762 12,955,453 43,077,426 3,066,822 4,405 99,842,612 3,419 77,477,868
Elevation 2140.00 9,621,848 27,341,372 12,076,400 42,485,149 2,831,880 4,829 62,320,722 3,747 48,360,880
Elevation 2130.00 9,622,366 37,335,117 12,732,713 42,447,296 2,925,688 6,630 73,925,020 5,145 57,365,816
Elevation 2120.00 9,625,715 31,523,737 13,695,741 43,015,405 3,210,568 5,473 100,500,011 4,247 77,988,008
Elevation 2110.00 9,624,093 32,208,151 12,806,872 42,465,159 2,980,376 5,706 68,037,942 4,428 52,797,443
Elevation 2100.00 9,622,622 34,179,284 13,324,379 42,667,702 3,096,834 6,030 79,940,457 4,679 62,033,795
Elevation 2090.00 9,619,555 28,624,330 12,140,447 42,441,936 2,838,426 5,076 58,845,125 3,939 45,663,817
Elevation 2080.00 9,628,918 35,090,334 12,474,049 42,905,364 2,877,643 6,194 81,031,058 4,806 62,880,101
Elevation 2070.00 9,624,630 28,993,026 12,941,047 42,857,911 3,038,491 5,056 85,560,266 3,924 66,394,767
Elevation 2060.00 9,621,541 28,384,376 9,916,674 42,385,513 2,283,789 5,029 59,880,729 3,902 46,467,445
Elevation 2050.00 9,662,571 31,925,895 13,218,188 43,153,065 3,086,480 5,611 81,156,842 4,354 62,977,709
Elevation 2040.00 9,654,058 29,830,005 12,213,544 42,428,700 2,849,336 5,252 72,809,496 4,076 56,500,169
Elevation 2030.00 9,641,288 35,746,696 13,418,225 42,540,906 3,112,821 6,181 121,739,848 4,796 94,470,122
Elevation 2020.00 9,536,663 34,600,255 13,879,150 42,197,072 3,232,736 6,104 81,002,611 4,736 62,858,026
Elevation 2010.00 9,528,478 31,923,163 13,393,076 42,227,111 3,129,136 5,652 68,483,387 4,386 53,143,108
Elevation 2000.00 9,522,661 39,475,263 15,488,844 42,564,472 3,602,915 6,918 106,307,095 5,368 82,494,306
Elevation 1990.00 9,457,766 46,406,926 14,266,289 41,829,155 3,245,968 8,249 89,671,195 6,401 69,584,848
Elevation 1980.00 9,445,909 51,162,105 15,682,269 42,139,121 3,572,942 8,878 164,663,453 6,889 127,778,840
Elevation 1970.00 9,435,226 47,174,842 15,105,093 41,960,335 3,453,978 8,256 130,337,038 6,407 101,141,542
Elevation 1960.00 9,429,946 37,106,416 14,217,274 41,880,778 3,296,890 6,659 52,430,757 5,167 40,686,267
Elevation 1950.00 9,446,502 59,257,392 15,041,979 41,943,059 3,355,443 10,346 171,342,763 8,028 132,961,984
Elevation 1940.00 9,433,090 43,164,730 14,330,607 41,741,437 3,288,949 7,515 131,437,446 5,831 101,995,459
Elevation 1930.00 9,415,318 33,261,814 12,630,563 41,510,774 2,931,694 5,790 101,476,235 4,493 78,745,558
Elevation 1920.00 9,356,635 36,287,687 10,835,532 41,065,834 2,467,573 6,100 176,655,762 4,734 137,084,872
Elevation 1910.00 9,334,368 42,349,766 11,805,429 40,835,705 2,675,339 6,904 271,436,265 5,358 210,634,541
Elevation 1900.00 9,267,184 38,443,692 9,756,401 40,006,008 2,192,710 6,105 295,815,803 4,738 229,553,064
Elevation 1890.00 9,242,189 30,992,846 8,230,910 39,504,916 1,862,798 4,788 279,321,626 3,715 216,753,582
Elevation 1880.00 9,215,891 22,904,176 6,850,172 39,060,596 1,567,299 3,589 191,090,043 2,785 148,285,873
Elevation 1870.00 9,181,584 18,006,565 5,545,494 38,514,810 1,269,906 2,907 124,119,545 2,256 96,316,767
Elevation 1860.00 9,147,789 15,757,662 4,898,402 37,961,342 1,121,509 2,588 95,320,754 2,008 73,968,905
Elevation 1850.00 9,123,883 21,915,101 4,500,830 37,727,449 978,819 3,691 104,494,575 2,864 81,087,790

316
Elevation 1840.00 9,106,318 17,010,375 3,941,386 37,493,496 872,562 2,838 89,443,509 2,202 69,408,163
Elevation 1830.00 9,090,488 12,593,760 3,422,890 37,303,833 773,824 2,046 82,995,508 1,587 64,404,514
Elevation 1820.00 9,078,981 7,194,576 2,678,609 37,060,244 619,413 1,303 6,684,634 1,011 5,187,276
Elevation 1810.00 9,071,219 6,688,723 2,205,226 36,894,177 504,733 1,204 8,440,405 934.0113 6,549,754
Elevation 1800.00 9,062,357 7,245,690 1,560,840 36,690,390 342,550 1,194 42,702,893 926.2184 33,137,445
Elevation 1790.00 9,056,654 2,567,146 1,098,557 36,539,766 257,866 423.285 15,007,978 328.4691 11,646,191
Elevation 1780.00 9,051,291 2,219,637 679,185 36,391,130 154,971 375.6501 10,035,073 291.5045 7,787,217
Elevation 1770.00 9,046,488 1,163,588 571,277 36,348,010 135,048 198.6992 4,720,615 154.1906 3,663,197
Elevation 1760.00 9,043,013 954,156 470,781 36,305,747 111,413 159.6237 4,879,002 123.868 3,786,105
Elevation 1750.00 9,039,635 684,504 367,443 36,263,473 87,443 111.4461 4,433,501 86.4822 3,440,397
Elevation 1740.00 9,036,195 533,661 295,403 36,229,341 70,467 84.729 4,113,223 65.7497 3,191,861
Elevation 1730.00 9,033,666 413,750 236,153 36,202,364 56,418 65.6459 3,202,658 50.9412 2,485,262
Elevation 1720.00 9,031,310 304,077 184,805 36,178,366 44,270 48.5246 2,268,665 37.6551 1,760,484
Elevation 1710.00 9,029,592 235,018 146,046 36,160,358 34,992 38.5344 1,439,899 29.9027 1,117,362
Elevation 1700.00 9,028,008 171,248 106,376 36,142,618 25,488 28.0368 1,061,847 21.7566 823,994
Elevation 1690.00 9,027,216 139,337 86,544 36,133,747 20,736 22.8096 864,816 17.7002 671,098
Elevation 1680.00 9,023,760 72,660 45,076 36,108,000 10,800 11.88 455,386 9.2189 353,379
Elevation 1670.00 9,023,760 0 0 36,095,040 0 0 0 0 0
Elevation 1660.00 9,023,760 0 0 36,095,040 0 0 0 0 0
Elevation 1650.00 9,023,760 0 0 36,095,040 0 0 0 0 0
Elevation 1640.00 9,023,760 0 0 36,095,040 0 0 0 0 0
Elevation 1630.00 9,023,760 0 0 36,095,040 0 0 0 0 0
Elevation 1620.00 9,023,760 0 0 36,095,040 0 0 0 0 0
Elevation 1610.00 9,023,760 0 0 36,095,040 0 0 0 0 0
Elevation 1600.00 9,023,760 0 0 36,095,040 0 0 0 0 0
Elevation 1590.00 9,023,760 0 0 36,095,040 0 0 0 0 0
Elevation 1580.00 9,023,760 0 0 36,095,040 0 0 0 0 0
Total 734,934,618 1,433,650,948 536,892,565 3,118,312,984 124,521,903 248,105 4,815,385,183 192,529 3,736,738,904

317
Fig. F. 1. 2D economic model (Section level 1970 m, Plan view)

318
Fig. F. 2. 2D economic model (Section level 1900 m, Plan View)

319
Fig. F. 3. 2D economic model (Section 638400 E North-South View)

320
Fig. F. 4. 2D economic model (Section 3644688.00 N East-West View)

321
Appendix. G: The Kahang Ultimate Pit Limit

Fig. G. 1. 2D pit limit (Section level 2225 m, Plan View)

322
Fig. G. 2. 3D Ultimate pit limit view of the Kahang deposit including all boreholes (Section

5140.00 E)

323
 Appendix. H: Extraction Sequences (Nested Pits)

Table. H. 1. Pit optimisation: Extraction sequences for incremental and cumulative NPV
generated by NPV Scheduler software (sequence No. 92 specifies OES)

Incremental Data
Processing Total Strip
Point Profit Revenue Mining Cost NPV Total Rock Total Ore
Cost Waste Ratio
$ $ $ $ $ Tonnes Tonnes Tonnes
1 -302,720 0 0 302,720 -301,976 170,280 0 170,280 0
2 -302,720 0 0 302,720 -300,489 170,280 0 170,280 0
3 -302,720 0 0 302,720 -299,010 170,280 0 170,280 0
4 -302,720 0 0 302,720 -297,537 170,280 0 170,280 0
5 -137,860 202,453 37,593 302,720 -134,990 171,420 7,980 163,440 20.479
6 26,464 412,809 83,624 302,720 25,506 172,867 18,009 154,857 8.599
7 793,812 1,268,154 171,621 302,720 769,741 174,971 33,969 141,001 4.151
8 105,371 499,501 91,410 302,720 101,672 175,492 19,329 156,163 8.079
9 -134,013 213,817 45,109 302,720 -128,436 172,222 9,772 162,449 16.623
10 -73,933 298,893 70,106 302,720 -70,094 173,537 15,427 158,110 10.249
11 -302,720 0 0 302,720 -287,275 170,280 0 170,280 0
12 64,932 467,785 100,133 302,720 61,523 177,492 21,734 155,758 7.166
13 -137,780 214,213 49,272 302,720 -129,308 173,390 10,812 162,577 15.036
14 -302,720 0 0 302,720 -282,968 170,280 0 170,280 0
15 -302,720 0 0 302,720 -281,574 170,280 0 170,280 0
16 -26,561 435,028 158,868 302,720 -24,522 177,174 36,691 140,483 3.829
17 -302,720 0 0 302,720 -278,752 170,280 0 170,280 0
18 -302,720 0 0 302,720 -277,380 170,280 0 170,280 0
19 -214,493 183,810 95,583 302,720 -195,599 172,751 22,631 150,120 6.633
20 -252,914 132,560 82,753 302,720 -229,454 174,083 19,784 154,299 7.799
21 -257,887 112,450 67,617 302,720 -232,801 173,248 16,134 157,113 9.738
22 -85,507 439,206 221,992 302,720 -77,039 192,001 52,473 139,528 2.659
23 -96,224 304,941 98,445 302,720 -86,066 174,462 22,491 151,971 6.757
24 -66,359 390,875 154,514 302,720 -59,158 178,339 35,945 142,393 3.961
25 -19,790 483,914 200,984 302,720 -17,443 179,244 46,889 132,355 2.823
26 21,957 567,760 243,083 302,720 19,451 184,651 56,820 127,830 2.25
27 -147,140 483,294 327,714 302,720 -128,609 186,587 78,619 107,968 1.373
28 -98,236 463,734 259,250 302,720 -85,557 182,576 61,686 120,889 1.96
29 -115,813 455,952 269,045 302,720 -100,337 179,969 64,195 115,774 1.803
30 -210,706 186,391 94,377 302,720 -181,214 177,382 22,313 155,068 6.95
31 -302,720 0 0 302,720 -259,296 170,280 0 170,280 0
32 -302,720 0 0 302,720 -258,020 170,280 0 170,280 0
33 -302,720 0 0 302,720 -256,749 170,280 0 170,280 0
34 -302,720 0 0 302,720 -255,484 170,584 0 170,584 0

324
35 -302,720 0 0 302,720 -254,225 170,280 0 170,280 0
36 -302,720 0 0 302,720 -252,973 170,280 0 170,280 0
37 -302,720 0 0 302,720 -251,727 170,280 0 170,280 0
38 -302,720 0 0 302,720 -250,488 170,280 0 170,280 0
39 -302,720 0 0 302,720 -249,254 170,280 0 170,280 0
40 -302,720 0 0 302,720 -248,027 170,280 0 170,280 0
41 -302,720 0 0 302,720 -246,806 170,280 0 170,280 0
42 -302,720 0 0 302,720 -245,590 170,280 0 170,280 0
43 -302,720 0 0 302,720 -244,381 170,280 0 170,280 0
44 -302,720 0 0 302,720 -243,177 170,705 0 170,705 0
45 -302,720 0 0 302,720 -241,974 170,948 0 170,948 0
46 -302,720 0 0 302,720 -240,781 170,280 0 170,280 0
47 -302,720 0 0 302,720 -239,595 170,280 0 170,280 0
48 -302,720 0 0 302,720 -238,416 170,280 0 170,280 0
49 -302,720 0 0 302,720 -237,242 170,280 0 170,280 0
50 -302,720 0 0 302,720 -236,074 170,280 0 170,280 0
51 -302,720 0 0 302,720 -234,911 170,280 0 170,280 0
52 -291,811 12,710 1,801 302,720 -225,335 170,280 360 169,920 472
53 -90,719 310,050 98,048 302,720 -69,561 170,280 22,320 147,960 6.629
54 -294,195 14,699 6,174 302,720 -224,927 170,280 1,440 168,840 117.25
55 -274,259 50,029 21,568 302,720 -208,641 170,280 5,040 165,240 32.786
56 -302,720 0 0 302,720 -229,169 177,332 0 177,332 0
57 -302,720 0 0 302,720 -228,008 170,401 0 170,401 0
58 -291,174 23,728 12,182 302,720 -218,235 170,523 2,880 167,643 58.209
59 -302,427 2,029 1,735 302,720 -225,540 176,907 420 176,486 419.407
60 -290,988 117,539 105,806 302,720 -215,870 183,144 25,688 157,456 6.13
61 -283,584 38,944 19,807 302,720 -209,301 170,462 4,680 165,782 35.424
62 -222,790 105,973 26,043 302,720 -163,611 170,888 5,760 165,128 28.668
63 -78,237 336,201 111,717 302,720 -57,438 176,603 25,560 151,043 5.909
64 58,983 558,886 197,182 302,720 43,221 170,948 45,360 125,588 2.769
65 -3,251 510,359 210,889 302,720 -2,364 175,132 49,145 125,987 2.564
66 177,601 853,272 372,951 302,720 127,686 175,630 87,265 88,364 1.013
67 113,192 776,331 360,418 302,720 81,061 179,400 84,694 94,705 1.118
68 -209,846 168,892 76,018 302,720 -149,388 180,129 17,832 162,297 9.101
69 406,831 1,080,798 371,247 302,720 288,339 176,116 85,246 90,870 1.066
70 251,995 721,505 166,789 302,720 178,042 177,089 36,608 140,481 3.837
71 -150,600 223,021 70,900 302,720 -105,686 187,547 16,163 171,384 10.603
72 -243,195 87,166 27,640 302,720 -169,701 188,144 6,300 181,843 28.86
73 -247,876 169,045 114,200 302,720 -172,022 176,707 27,430 149,276 5.442
74 129,464 629,298 197,114 302,720 89,151 178,622 44,865 133,756 2.981
75 92,462 621,500 226,317 302,720 63,619 175,950 52,219 123,731 2.369
76 1,112,962 1,882,716 467,033 302,720 760,176 176,907 103,460 73,446 0.71
77 1,589,793 2,450,295 557,781 302,720 1,080,859 173,441 122,120 51,321 0.42

325
 78 1,263,254 2,087,980 522,005 302,720 854,637 180,251 115,748 64,502 0.557
79 1,541,488 2,339,445 495,237 302,720 1,036,973 172,772 107,238 65,534 0.611
80 1,578,076 2,398,261 517,465 302,720 1,055,895 180,731 112,385 68,345 0.608
81 1,528,337 2,422,321 591,264 302,720 1,017,190 185,284 130,700 54,584 0.418
82 2,781,670 3,698,135 613,744 302,720 1,842,401 182,654 127,238 55,416 0.436
83 1,465,090 2,214,107 446,296 302,720 965,308 181,782 95,931 85,851 0.895
84 1,210,660 2,057,535 544,154 302,720 793,814 186,803 121,572 65,230 0.537
85 388,737 1,214,670 523,213 302,720 253,478 187,854 122,406 65,447 0.535
86 737,539 1,691,310 651,051 302,720 477,955 190,339 151,100 39,238 0.26
87 391,134 1,261,510 567,655 302,720 252,373 191,819 133,255 58,563 0.439
88 296,330 1,010,935 411,885 302,720 189,996 186,951 96,048 90,903 0.946
89 474,331 1,278,375 501,323 302,720 301,919 190,363 116,561 73,801 0.633
90 808,655 1,580,718 469,343 302,720 512,737 191,439 106,399 85,039 0.799
91 1,747,427 2,596,472 546,324 302,720 1,101,455 190,678 118,568 72,110 0.608
92 1,732,457 2,464,485 429,308 302,720 1,086,318 190,430 90,212 100,217 1.111
93 3,810 460,839 154,308 302,720 2,278 179,721 35,340 144,380 4.085
94 -99,673 431,512 228,464 302,720 -61,923 182,126 54,153 127,972 2.363
95 151,539 643,635 189,376 302,720 93,866 181,168 42,872 138,296 3.226
96 -219,275 250,807 167,361 302,720 -134,725 178,635 40,136 138,499 3.451
97 10,026 517,422 204,675 302,720 6,124 188,830 47,606 141,224 2.966
98 -7,950 443,913 149,143 302,720 -4,579 184,275 34,216 150,059 4.386
99 114,108 642,500 225,672 302,720 68,995 181,112 51,915 129,196 2.489
100 296,279 774,060 221,780 256,000 177,855 150,406 50,110 100,296 2.002
Cumulative Data

Processing Total Strip

Point Profit Revenue Mining Cost NPV Total Rock Total Ore
Cost Waste Ratio
$ $ $ $ $ tonnes tonnes tonnes
1 -302,720 0 0 302,720 -301,976 170,280 0 170,280 0
2 -605,440 0 0 605,440 -602,465 340,560 0 340,560 0
3 -908,160 0 0 908,160 -901,474 510,840 0 510,840 0
4 -1,210,880 0 0 1,210,880 -1,199,011 681,120 0 681,120 0
5 -1,348,740 202,453 37,593 1,513,600 -1,334,000 852,540 7,980 844,560 105.824
6 -1,322,276 615,262 121,218 1,816,320 -1,308,493 1,025,408 25,990 999,417 38.453
7 -528,463 1,883,417 292,839 2,119,040 -538,752 1,200,379 59,959 1,140,419 19.02
8 -423,092 2,382,918 384,250 2,421,760 -437,079 1,375,871 79,289 1,296,582 16.352
9 -557,105 2,596,735 429,359 2,724,480 -565,515 1,548,094 89,062 1,459,031 16.382
10 -631,037 2,895,629 499,466 3,027,200 -635,609 1,721,631 104,489 1,617,142 15.477
11 -933,757 2,895,629 499,466 3,329,920 -922,883 1,891,911 104,489 1,787,422 17.106
12 -868,825 3,363,415 599,599 3,632,640 -861,360 2,069,404 126,223 1,943,180 15.395
13 -1,006,604 3,577,628 648,872 3,935,360 -990,667 2,242,795 137,036 2,105,758 15.366
14 -1,309,324 3,577,628 648,872 4,238,080 -1,273,634 2,413,075 137,036 2,276,038 16.609
15 -1,612,044 3,577,628 648,872 4,540,800 -1,555,208 2,583,355 137,036 2,446,318 17.852

326
16 -1,638,605 4,012,656 807,740 4,843,520 -1,579,729 2,760,529 173,727 2,586,801 14.89
17 -1,941,325 4,012,656 807,740 5,146,240 -1,858,481 2,930,809 173,727 2,757,081 15.87
18 -2,244,045 4,012,656 807,740 5,448,960 -2,135,860 3,101,089 173,727 2,927,361 16.85
19 -2,458,538 4,196,467 903,324 5,751,680 -2,331,459 3,273,841 196,359 3,077,481 15.673
20 -2,711,451 4,329,027 986,077 6,054,400 -2,560,913 3,447,924 216,143 3,231,780 14.952
21 -2,969,337 4,441,478 1,053,694 6,357,120 -2,793,714 3,621,172 232,278 3,388,894 14.59
22 -3,054,844 4,880,684 1,275,687 6,659,840 -2,870,752 3,813,174 284,751 3,528,422 12.391
23 -3,151,067 5,185,625 1,374,132 6,962,560 -2,956,817 3,987,636 307,243 3,680,393 11.979
24 -3,217,426 5,576,501 1,528,646 7,265,280 -3,015,975 4,165,976 343,188 3,822,787 11.139
25 -3,237,216 6,060,415 1,729,631 7,568,000 -3,033,417 4,345,220 390,078 3,955,142 10.139
26 -3,215,259 6,628,176 1,972,714 7,870,720 -3,013,965 4,529,872 446,899 4,082,972 9.136
27 -3,362,398 7,111,471 2,300,429 8,173,440 -3,142,574 4,716,459 525,518 4,190,940 7.975
28 -3,460,634 7,575,206 2,559,679 8,476,160 -3,228,131 4,899,035 587,204 4,311,830 7.343
29 -3,576,446 8,031,158 2,828,724 8,778,880 -3,328,467 5,079,004 651,400 4,427,604 6.797
30 -3,787,152 8,217,550 2,923,102 9,081,600 -3,509,681 5,256,387 673,713 4,582,673 6.802
31 -4,089,872 8,217,550 2,923,102 9,384,320 -3,768,976 5,426,667 673,713 4,752,953 7.055
32 -4,392,592 8,217,550 2,923,102 9,687,040 -4,026,996 5,596,947 673,713 4,923,233 7.308
33 -4,695,312 8,217,550 2,923,102 9,989,760 -4,283,744 5,767,227 673,713 5,093,513 7.56
34 -4,998,032 8,217,550 2,923,102 10,292,480 -4,539,228 5,937,811 673,713 5,264,097 7.814
35 -5,300,752 8,217,550 2,923,102 10,595,200 -4,793,452 6,108,091 673,713 5,434,377 8.066
36 -5,603,472 8,217,550 2,923,102 10,897,920 -5,046,424 6,278,371 673,713 5,604,657 8.319
37 -5,906,192 8,217,550 2,923,102 11,200,640 -5,298,151 6,448,651 673,713 5,774,937 8.572
38 -6,208,912 8,217,550 2,923,102 11,503,360 -5,548,638 6,618,931 673,713 5,945,217 8.825
39 -6,511,632 8,217,550 2,923,102 11,806,080 -5,797,892 6,789,211 673,713 6,115,497 9.077
40 -6,814,352 8,217,550 2,923,102 12,108,800 -6,045,918 6,959,491 673,713 6,285,777 9.33
41 -7,117,072 8,217,550 2,923,102 12,411,520 -6,292,723 7,129,771 673,713 6,456,057 9.583
42 -7,419,792 8,217,550 2,923,102 12,714,240 -6,538,313 7,300,051 673,713 6,626,337 9.836
43 -7,722,512 8,217,550 2,923,102 13,016,960 -6,782,694 7,470,331 673,713 6,796,617 10.088
44 -8,025,232 8,217,550 2,923,102 13,319,680 -7,025,870 7,641,036 673,713 6,967,323 10.342
45 -8,327,952 8,217,550 2,923,102 13,622,400 -7,267,843 7,811,985 673,713 7,138,272 10.595
46 -8,630,672 8,217,550 2,923,102 13,925,120 -7,508,623 7,982,265 673,713 7,308,552 10.848
47 -8,933,392 8,217,550 2,923,102 14,227,840 -7,748,218 8,152,545 673,713 7,478,832 11.101
48 -9,236,112 8,217,550 2,923,102 14,530,560 -7,986,634 8,322,825 673,713 7,649,112 11.354
49 -9,538,832 8,217,550 2,923,102 14,833,280 -8,223,875 8,493,105 673,713 7,819,392 11.606
50 -9,841,552 8,217,550 2,923,102 15,136,000 -8,459,948 8,663,385 673,713 7,989,672 11.859
51 -10,144,272 8,217,550 2,923,102 15,438,720 -8,694,858 8,833,665 673,713 8,159,952 12.112
52 -10,436,083 8,230,261 2,924,903 15,741,440 -8,920,193 9,003,945 674,073 8,329,872 12.358
53 -10,526,801 8,540,311 3,022,952 16,044,160 -8,989,753 9,174,225 696,393 8,477,832 12.174
54 -10,820,996 8,555,010 3,029,126 16,346,880 -9,214,680 9,344,505 697,833 8,646,672 12.391
55 -11,095,255 8,605,040 3,050,694 16,649,600 -9,423,320 9,514,785 702,873 8,811,912 12.537
56 -11,397,975 8,605,040 3,050,694 16,952,320 -9,652,489 9,692,118 702,873 8,989,244 12.789
57 -11,700,695 8,605,040 3,050,694 17,255,040 -9,880,497 9,862,520 702,873 9,159,646 13.032
58 -11,991,868 8,628,768 3,062,876 17,557,760 -10,098,731 10,033,043 705,753 9,327,289 13.216

327
59 -12,294,295 8,630,798 3,064,612 17,860,480 -10,324,271 10,209,950 706,174 9,503,776 13.458
60 -12,585,282 8,748,337 3,170,418 18,163,200 -10,540,140 10,393,094 731,862 9,661,232 13.201
61 -12,868,865 8,787,281 3,190,226 18,465,920 -10,749,441 10,563,556 736,542 9,827,014 13.342
62 -13,091,655 8,893,255 3,216,269 18,768,640 -10,913,051 10,734,444 742,302 9,992,142 13.461
63 -13,169,891 9,229,456 3,327,987 19,071,360 -10,970,489 10,911,048 767,862 10,143,185 13.21
64 -13,110,908 9,788,343 3,525,170 19,374,080 -10,927,267 11,081,996 813,222 10,268,774 12.627
65 -13,114,158 10,298,702 3,736,059 19,676,800 -10,929,631 11,257,129 862,368 10,394,761 12.054
66 -12,936,557 11,151,975 4,109,011 19,979,520 -10,801,944 11,432,760 949,633 10,483,126 11.039
67 -12,823,364 11,928,307 4,469,430 20,282,240 -10,720,883 11,612,160 1,034,328 10,577,832 10.227
68 -13,033,209 12,097,199 4,545,448 20,584,960 -10,870,270 11,792,289 1,052,160 10,740,129 10.208
69 -12,626,378 13,177,998 4,916,696 20,887,680 -10,581,930 11,968,406 1,137,406 10,831,000 9.523
70 -12,374,382 13,899,504 5,083,486 21,190,400 -10,403,888 12,145,496 1,174,014 10,971,481 9.345
71 -12,524,982 14,122,525 5,154,386 21,493,120 -10,509,573 12,333,043 1,190,177 11,142,865 9.362
72 -12,768,176 14,209,692 5,182,027 21,795,840 -10,679,274 12,521,187 1,196,478 11,324,708 9.465
73 -13,016,051 14,378,738 5,296,228 22,098,560 -10,851,295 12,697,894 1,223,908 11,473,985 9.375
74 -12,886,587 15,008,036 5,493,342 22,401,280 -10,762,144 12,876,516 1,268,774 11,607,742 9.149
75 -12,794,124 15,629,537 5,719,660 22,704,000 -10,698,524 13,052,467 1,320,993 11,731,473 8.881
76 -11,681,161 17,512,253 6,186,694 23,006,720 -9,938,348 13,229,374 1,424,454 11,804,920 8.287
77 -10,091,368 19,962,549 6,744,476 23,309,440 -8,857,489 13,402,816 1,546,574 11,856,241 7.666
78 -8,828,113 22,050,529 7,266,481 23,612,160 -8,002,851 13,583,067 1,662,323 11,920,744 7.171
79 -7,286,625 24,389,974 7,761,718 23,914,880 -6,965,878 13,755,840 1,769,561 11,986,278 6.774
80 -5,708,548 26,788,236 8,279,184 24,217,600 -5,909,983 13,936,571 1,881,947 12,054,624 6.405
81 -4,180,211 29,210,558 8,870,448 24,520,320 -4,892,792 14,121,856 2,012,648 12,109,208 6.017
82 -1,398,540 32,908,693 9,484,193 24,823,040 -3,050,390 14,304,510 2,139,886 12,164,624 5.685
83 66,550 35,122,800 9,930,489 25,125,760 -2,085,082 14,486,292 2,235,817 12,250,475 5.479
84 1,277,211 37,180,336 10,474,644 25,428,480 -1,291,268 14,673,096 2,357,390 12,315,705 5.224
85 1,665,948 38,395,007 10,997,858 25,731,200 -1,037,789 14,860,950 2,479,796 12,381,153 4.993
86 2,403,488 40,086,317 11,648,909 26,033,920 -559,834 15,051,289 2,630,897 12,420,392 4.721
87 2,794,623 41,347,827 12,216,564 26,336,640 -307,461 15,243,108 2,764,153 12,478,955 4.515
88 3,090,953 42,358,763 12,628,450 26,639,360 -117,464 15,430,060 2,860,201 12,569,859 4.395
89 3,565,284 43,637,138 13,129,773 26,942,080 184,456 15,620,424 2,976,763 12,643,660 4.247
90 4,373,939 45,217,857 13,599,117 27,244,800 697,194 15,811,864 3,083,163 12,728,700 4.128
91 6,121,367 47,814,330 14,145,442 27,547,520 1,798,650 16,002,542 3,201,731 12,800,811 3.998
92 7,853,825 50,278,816 14,574,750 27,850,240 2,884,968 16,192,972 3,291,944 12,901,028 3.919
93 7,857,636 50,739,655 14,729,058 28,152,960 2,887,246 16,372,694 3,327,284 13,045,409 3.921
94 7,757,963 51,171,167 14,957,523 28,455,680 2,825,323 16,554,820 3,381,438 13,173,382 3.896
95 7,909,502 51,814,802 15,146,900 28,758,400 2,919,189 16,735,988 3,424,310 13,311,678 3.887
96 7,690,228 52,065,610 15,314,261 29,061,120 2,784,465 16,914,624 3,464,446 13,450,177 3.882
97 7,700,255 52,583,032 15,518,937 29,363,840 2,790,590 17,103,454 3,512,052 13,591,401 3.87
98 7,692,305 53,026,946 15,668,080 29,666,560 2,786,011 17,287,729 3,546,268 13,741,460 3.875
99 7,806,413 53,669,446 15,893,753 29,969,280 2,855,007 17,468,841 3,598,184 13,870,657 3.855
100 8,102,692 54,443,507 16,115,534 30,225,280 3,032,862 17,619,248 3,648,294 13,970,953 3.829

328
329